From Agglomeration to Innovation: Upgrading Industrial Clusters in Emerging Economies

From Agglomeration to Innovation

Other titles from IDE-JETRO: MAKING HEALTH SERVICES MORE ACCESSIBLE IN DEVELOPING COUNTRIES Hiroko Uchimura (editor) GLOBALISATION, EMPLOYMENT AND MOBILITY The South Asian Experience Hiroshi Sato and Mayumi Murayama (editors) POVERTY, REDUCTION AND BEYOND Development Strategies for Low-Income Countries Takashi Shiraishi, Tatsufumi Yamagata and Shahid Yusuf (editors) RECOVERING FINANCIAL SYSTEMS China and Asian Transition Economies Mariko Watanabe (editor) EAST ASIA’S DE FACTO ECONOMIC INTEGRATION Daisuke Hiratsuka (editor) NEW DEVELOPMENTS OF THE EXCHANGE RATE REGIMES IN DEVELOPING COUNTRIES Hisayuki Mitsuo (editor) DEVELOPMENT OF ENVIRONMENTAL POLICY IN JAPAN AND ASIAN COUNTRIES Tadayoshi Terao and Kenji Otsuka (editors) ECONOMIC INTEGRATION IN ASIA AND INDIA Masahisa Fujita (editor) REGIONAL INTEGRATION IN EAST ASIA From the Viewpoint of Spatial Economics Masahisa Fujita (editor) INDUSTRIAL CLUSTERS IN ASIA Analyses of Their Competition and Cooperation Akifumi Kuchiki and Masatsugu Tsuji (editors) GENDER AND DEVELOPMENT The Japanese Experience in Comparative Perspective Mayumi Murayama (editor) SPATIAL STRUCTURE AND REGIONAL DEVELOPMENT IN CHINA An Interregional Input-Output Approach Nobuhiro Okamoto and Takeo Ihara (editors) THE FLOWCHART APPROACH TO INDUSTRIAL CLUSTER POLICY Akifumi Kuchiki and Masatsugu Tsuji (editors) FROM AGGLOMERATION TO INNOVATION Upgrading Industrial Clusters in Emerging Economies Akifumi Kuchiki and Masatsugu Tsuji (editors)

From Agglomeration to Innovation Upgrading Industrial Clusters in Emerging Economies

Edited by Akifumi Kuchiki and Masatsugu Tsuji

© Institute of Developing Economies (IDE), JETRO 2010 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6-10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2010 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN: 978–0–230–23310–2 hardback This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. 10 9 8 7 6 5 4 3 2 1 19 18 17 16 15 14 13 12 11 10 Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne

Contents List of Tables

vii

List of Figures

xi

Acknowledgements

xiii

Notes on the Contributors

xiv

1 Introduction Masatsugu Tsuji and Akifumi Kuchiki Part I

1

Learning Linkage as Drivers for Creating Cluster and Innovation

2 The Automobile Industry Cluster in Malaysia Akifumi Kuchiki

15

3 Industrial Cluster Development and Innovation in Singapore Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

50

4 Empirical Analysis of the Relationship between Upgrading and Innovation of Japanese SMEs and Industrial Clustering Shoichi Miyahara and Masatsugu Tsuji 5 Collective Goods for Reformatting the Rio de Janeiro Software Cluster into a Local Innovation System Antonio José Junqueira Botelho, Alex da Silva Alves and Glaudson Mosqueira Bastos 6 Innovation through Long-distance Conversations? Experience from Offshoring-based Software Clusters in Bangalore, India Aya Okada Part II

117

166

204

Market Pooling and Organizational Change as Drivers for Creating Cluster and Innovation

7 A Comparative Analysis of Organizational Innovation in Japanese SMEs Generated by Information Communication Technology Masatsugu Tsuji and Shoichi Miyahara

v

231

vi

Contents

8 The Role of the Specialized Markets in Upgrading Industrial Clusters in China Ke Ding 9 Industrial Clusters and Workplace Training to Expand Innovation Capability: Evidence from Manufacturing in the Greater Bangkok, Thailand Tomohiro Machikita 10 Innovation as a Driver for Building an Oil & Gas Industrial Cluster in Rio de Janeiro, Brazil Antonio José Junqueira Botelho and Glaudson Mosqueira Bastos

270

290

326

11 Conclusion Akifumi Kuchiki and Masatsugu Tsuji

357

Index

361

Tables 2.1 Changes in countries and regions for promising businesses in the medium term 2.2 Suitable production sites in the medium and long term 2.3 Problems in localization of employees, products and technology 2.4 Reasons for promising countries and regions 2.5 Evaluation index of investment environment of ASEAN and India in comparison with China 2.6 Questionnaire survey on application of Flowchart Approach to industrial cluster policy 3.1a Profile of the Singapore pharmaceuticals sector, 1980–2006 3.1b Profile of the Singapore medical technology sector, 1980–2006 3.1c Profile of the Singapore biomedical sciences (BMS) sector, 1980–2006 3.2 Milestones in the Singapore biomedical sector 3.3 Pharmaceutical and medical technology share of Singapore biomedical sector, 1980–2006 3.4 R&D expenditure and manpower in the biomedical sector, 1993–2006 3.5 Biomedical Shares of Singapore R&D Expenditure and RSEs, 1993–2006 3.6 Share of life science patents in Singapore, 1977–2007 3.7 Breakdown of Singapore life science patents by assignee, 1977–2007 3.8 Top pharmaceutical companies in Singapore, 2005 3.9 Major foreign pharmaceutical companies operating in Singapore 3.10 Establishment of life science public research institutes under A*STAR 3.11 Dedicated biotechnology firms (DBFs) founded in Singapore 3.12 Profile of NUS biomedical-related spin-off companies 3.13 Key component industries within Singapore maritime cluster 3.14 Growth trends in Singapore maritime cluster 3.15 Maritime clusters value added, international benchmarks 2001 3.16 Linkages of maritime cluster to economy 3.17 Principal statistics of Singapore maritime sector, 2005 3.18 Sales revenue of marine and offshore engineering industry 3.19 Principal statistics of marine and offshore engineering industry vii

30 31 33 34 35 37 56 57 58 59 61 62 63 63 64 66 67 73 77 80 86 87 87 88 89 95 97

viii

Tables

3.20 Leading offshore engineering companies 3.21 Leading offshore support services companies 3.22 Key R&D indicators for the Singapore marine engineering sector 1993–2006 3.23 Offshore patents invented in Singapore or assigned to Singapore interests 3.24 Examples of private and public/IHE collaborations in offshore sector 3.25 Profile of Keppel FELS and Sembcorp Marine 3.26 Turnover and net profit for Keppel O&M Ltd and Sembcorp Marine Ltd, 1993–2005 4.1 Distance between SMEs and collaborating partners 4.2 Year of establishment 4.3 Amount of capital 4.4 Number of employment 4.5 Category of industry 4.6 Category of manufacturing 4.7 Subcontracting 4.8 Recent annual sales 4.9 Trend of sales amount within recent 3 years 4.10 Balance of revenues and costs in recent 3 years 4.11 Ratio of R&D expenditures to total sales 4.12 Year of authorization 4.13 Number of upgrading and innovation: Replies to question V 4.14 Number of patents applied for 4.15 Number of patents registered 4.16 Number of new products and services developed 4.17 Ratio of R&D and sales trend 4.18 Ratio of R&D and business performance 4.19 Summary of statistics 4.20 Summary of estimation results: Upgrading model 4.21 Summary of estimation results: Innovation model A1–1 Results of estimation: Upgrading model I A1–2 Results of estimation: Upgrading model II A1–3 Results of estimation: Upgrading model III A1–4 Results of estimation: Upgrading model IV A2–1 Results of estimation: Innovation model I A2–2 Results of estimation: Innovation model II A2–3 Results of estimation: Innovation model III 5.1 Characteristics of the Rio de Janeiro City IT local productive arrangement 5.2 Local collective goods for the City of Rio de Janeiro 6.1 Analytic and interpretive perspectives 6.2 Modes of delivery of software services exports from India (%)

98 100 102 103 104 106 108 120 120 121 121 122 122 123 123 123 124 124 125 126 129 129 129 131 131 134 140 144 149 151 154 156 158 160 162 181 196 209 214

Tables ix

6.3 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11 7.12 7.13 8.1 8.2 8.3 8.4 8.5 9.1

9.2

9.3

9.4 9.5 9.6 9.7

9.8

TCS’s partnerships with universities worldwide Question on software use Question on Internet use Index of organizational innovation of two groups Result of component analysis Summary statistics Result of OLS estimation Factors affecting organizational innovation (1) Factors affecting organizational innovation (2) Probit/logit estimation Problems of organizational innovation by SMEs (1) Problems of organizational innovation by SMEs (2) Policy desired for organizational innovation (1) Policy desired for organizational innovation (2) Number of booths in specialized markets in Zhejiang’s major industrial clusters (1998) Scope of specialized markets in Zhejiang’s major industrial clusters (1998, multiple) Profile of the Yuyao Moulds cluster Raw material businesses in the Yuyao Market Division of labour in the Yuyao Market Summary statistics of training incidence by outside labour market experience and tenure for production and non-production workers: On-the-job training incidence in 2001 Summary statistics of training incidence by outside labour market experience and tenure for production and non-production workers: Off-the-job training incidence in 2001 Summary statistics of log of wage in July 2001 by outside labour market experience and tenure for production and non-production workers: Log of monthly wage Job tenure and previous experience by industry and occupation Industry effects on training incidence, dependent variable: Binomial OJT incidence in 2001 Industry effects on training incidence for production workers, dependent variable: Binomial OJT incidence in 2001 Industry effects on training incidence for non-production workers, dependent variable: Binomial OJT incidence in 2001 Industry effects on training incidence, dependent variable: Binomial OFFJT incidence in 2001

221 240 240 242 244 245 247 249 250 251 253 254 255 256 271 272 280 281 281

299

300

302 303 306 308

309 310

x Tables

9.9

9.10

9.11

9.12

9.13

9.14

9.15

9.16

10.1 10.2 10.3 10.4 10.5 10.6

Industry effects on training incidence for production workers, dependent variable: Binomial OFFJT incidence in 2001 Industry effects on training incidence for non-production workers, dependent variable: Binomial OFFJT incidence in 2001 Effects of interactions between OJT length and industry on log of wage, dependent variable: Log of wage in July 2001 Effects of interactions between OJT length and industry on log of wage for production workers, dependent variable: Log of wage in July 2001 Effects of interactions between OJT length and industry on log of wage for non-production workers, dependent variable: Log of wage in July 2001 Effects of interactions between OFFJT length and industry on log of wage, dependent variable: Log of wage in July 2001 Effects of interactions between OFFJT length and industry on log of wage for production workers, dependent variable: Log of wage in July 2001 Effects of interactions between OFFJT length and industry on log of wage for non-Production workers, dependent variable: Log of wage in July 2001 Evolution of O&G industry suppliers South-east region O&G industry suppliers O&G industry suppliers by service groups Oil proven reserves at end of 2006 Oil proven reserves at end of 2006 in South and Central America Sectoral funds: Regulatory frame and resources

311

312

313

315

316

317

319

320 327 327 328 329 329 338

Figures 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 3.1 3.2 3.3 3.4 4.1 5.1 6.1 6.2 7.1 7.2 7.3 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9 10.10 10.11

Flowchart Approach to industrial cluster policy Factors Cluster policy Actors Flowchart Approach: Step 1: Agglomeration Flowchart Approach: Step 2: Infrastructure Priorities of each player Number of firms to strengthen and enlarge their branches Prescriptions for the automobile industry cluster Prescriptions for automobile industry clustering in Malaysia Flowchart of innovation by university, industry and cluster The Biology industry cluster in Singapore Singapore’s BMS cluster development strategy Overall institutional framework in Singapore for IMC development Singapore’s IMC development strategy Shipbuilding and Repair Revenues, 1972–2006 Trend of Upgrading and Innovation A flowchart of IT cluster in the City of Rio de Janeiro Software development services life cycle by location of activities Pattern of intra-firm international division of labour: The case of TI’s semiconductor production Layer of questions in AHP Weight obtained by AHP Distributions of indices Brazil oil production and consumption Brazil proven reserves of oil and NGL at end of 2006 PETROBRAS evolution investments, 1991–2001 PETROBRAS technological cooperation programmes Regional distribution of PETROBRAS cooperation PETROBRAS deep sea drilling technology evolution PETROBRAS technological system PETROBRAS R&D expenditures PETROBRAS production targets Projected special participation expenditures on R&D Management structure of the programme (PROMINP) xi

18 19 19 19 20 21 23 30 39 41 44 46 65 91 92 96 128 170 215 219 241 241 243 330 330 332 340 341 343 344 345 346 347 348

xii Figures

10.12 10.13

Organizational structure of REDE PETRO BC A flowchart of the exploration and production (E&P) Segment of the O&G sector in the state of Rio de Janeiro

349 353

Acknowledgements First of all, the editors would like to express sincere gratitude to Dr Tomohiro Machikita, Dr Yasushi Ueki and Mr Kentaro Yoshida for their research input and comments during the whole process of this project; their encouragement and support have been invaluable. We are also grateful for lively discussions with Dr Keshab Das, Dr Toshitaka Gokan, Prof. Yoshiaki Hisamatsu, Mr Ikumo Isono, Dr Jobaid Kabir, Dr Hisaki Kono, Mr Kazuki Minato and Prof. Yumiko Okamoto. Special thanks are due to anonymous referees, who kindly provided valuable comments and suggestions for revising the draft. Finally, the editors are indebted to Ms Mariko Hashimoto, Ms Junko Yaegashi and Ms Mayumi Hasegawa for their generous secretarial works.

xiii

Contributors Akifumi Kuchiki Professor, Department of International Development Studies, College of Bioresource Sciences, Nihon University Masatsugu Tsuji Professor, Graduate School of Applied Informatics, University of Hyogo Alex da Silva Alves Research Fellow, Pontifícia Universidade Católica do Rio de Janeiro (PUC – Rio University), Brazil Glaudson Mosqueira Bastos Research Fellow, Pontifícia Universidade Católica do Rio de Janeiro (PUC – Rio University), Brazil Antonio José Junqueira Botelho Professor, Pontifícia Universidade Católica do Rio de Janeiro (PUC – Rio University), Brazil Ke Ding Research Fellow, Area Studies Center, IDE-JETRO Yuen-Ping Ho Research Manager, NUS Entrepreneurship Centre, National University of Singapore Tomohiro Machikita Research Fellow, Inter-disciplinary Studies Center, IDE-JETRO Shoichi Miyahara Professor, Faculty of Economics, Aoyama Gakuin University Aya Okada Professor, Graduate School of International Development, Nagoya University Annette Singh Research Officer, NUS Entrepreneurship Centre, National University of Singapore Poh-Kam Wong Professor, Business School and Lee Kuan Yew School of Public Policy, National University of Singapore

xiv

1 Introduction Masatsugu Tsuji and Akifumi Kuchiki

1.1 Background and previous studies 1.1.1 Background of this book Industrial agglomeration, which is now under way not only in East Asia but also in emerging economies such as India and Brazil, is at the centre of global attention. Thanks to agglomeration, these individual economies have been able to achieve economic growth, alleviate poverty and reduce regional gaps such as in income inequality. There is no doubt that global economies have also been enjoying economic growth through the increasing flow of products, funds, human resources and technologies. The interdependency and collaboration among global economies are expected to strengthen even further. In these circumstances, a variety of new structural shifts in these clusters is emerging. Agglomeration in East Asia was originally triggered by FDI (Foreign Direct Investment) by MNCs (multinational corporations), which aimed to establish alternative production bases outside their home economies in order to access relatively cheap natural resources, including unskilled labour, and engage in production related to raw materials and simple parts. The continuation of agglomeration, however, has resulted in upgrade their functions in East Asia and other emerging economies: (1) from being sub-contractors for simple work to producing intermediate goods; (2) from producing intermediate goods to final products; (3) from simple to complex or precision work; and (4) from low-value added to high-value added business activities. In other words, they have been upgrading from simple production related to raw materials and simple parts to production related to complex components and parts or related to higher R&D or innovation. This kind of upgrading becomes possible thanks to accumulated or transferred knowledge and know-how, in accordance with agglomeration. It can be said that these regions have crossed the threshold into the stage of knowledge-based economies. This type of industrial transformation from a production-based to a knowledge-based economy now occurring in East Asia and emerging 1

2

Masatsugu Tsuji and Akifumi Kuchiki

economies has yet to be fully analysed from the viewpoint of not only surveying the current situation but also establishing an economic theory to explain the transformation. This book attempts a forthright analysis of these topics, currently among the most interesting in the economic development field. Our goals are two-fold. First, we develop a simple but coherent model to postulate a hypothesis of the endogenous innovation processes; that is, the process of upgrading industrial clusters to the higher R&D or innovation stage, by viewing case studies in East Asia and other emerging economies. Second, we focus on the following specific issues: (1) the effects of adopting new technologies on the creation of product and process innovation in each industrial cluster; and (2) policy recommendations to foster the endogenous innovation process. Thus, we attempt to analyse to what extent our proposing hypothesis characterizes counterfactual evidence about ongoing transformation. Moreover, we use theoretical as well as econometric models to analyse the impact of public policy aiming to foster not only industrial agglomeration but also the innovation process. This enables us to grasp comparable characteristics for the respective industrial clusters and to formulate alternative policy recommendations. Finally, we attempt to explore ways the causal relationship between industrial agglomeration and fostering innovation systems can be connected each other by using evidence-based policymaking motivated by economic theory and convincing fact-finding. 1.1.2 Bottom line: Relationship to previous studies Innovation activities often occur at the local level, even if improvements in communication technologies expand information spillovers between cities and regions. Previous studies have explained this by emphasizing the role of local knowledge spillovers, labour turnover, spin-offs and entrepreneurship. In fact, Saxenian (1994), Porter (1998, 2000) and other associated researchers have produced landmark research studies in this field. Saxenian (1994) suggested that local, horizontal and flexible inter-firm networks among small and medium-sized enterprises (hereafter, SMEs) play an important role in fostering industrial clusters as well as innovation. Her work focuses not only on the role of local autonomy but also on vertically integrated inter-firm relations mostly governed by large enterprises. In the case of East Asia and other emerging economies, the situation is different. The relationship between clustering and incremental innovation seems to be quite different between cases of Saxenian (1994) and Porter (1998, 2000) on the one hand and of Asian industrial clusters on the other. Several studies have identified main driving forces that differentiate Asian and emerging economy-type clusters from Silicon Valley-type clusters in the context of innovation activities. Notwithstanding his use of less rigorous empirical analysis, Hashimoto (1997) emphasized that vertically integrated inter-firm cooperation is observed more commonly than horizontal inter-firm networks in industrial

Introduction 3

clusters in Japan. Recently, Arita et al. (2006) examined the evidence for regional cooperation within three Japanese clusters with a detailed empirical analysis. They found three important results: (1) positive correlations exist between the intensity of regional cooperation and corporate growth rate and R&D expenditures; (2) alliances with universities and cross-industry exchange organizations have significant positive effects on the growth rate of firms; (3) contents and partners of regional cooperation differ among the three clusters. These results are clearly different from those of Hashimoto (1997) who emphasized the role of vertically integrated inter-firm linkages in Japan. Needless to say, this topic requires the further accumulation of empirical results. This book uses the typology of industrial districts by Markusen (1996) to develop concrete ideas about extending the typology between Silicon Valleytype and East Asian or emerging economy-type clusters. Markusen’s work is also used to estimate the source of differences between the two types of clusters, and it is emphasized that although in the economic space information mobility is enhanced and the cost of information is decreased (slippery spaces), spaces suitable for knowledge creation are limited by the scarcity of resources and information required for innovation activities (sticky places). In our previous studies (Kuchiki and Tsuji 2005, 2008 and Tsuji et al. 2007), we extensively analysed the agglomeration process currently occurring in East Asia and postulated a hypothesis which is referred to as the ‘Flowchart Approach’. This postulates that MNCs, which are referred to as anchor firms, establish production bases first; followed by SMEs, which are suppliers or sub-contractors, and local firms establishing facilities near them. The Flowchart Approach also identifies factors which attract firms to particular regions, including: (1) natural resources such as raw materials and human resources, skilled labour and professional or unskilled labour; (2) physical infrastructure including highways, roads, airports, electricity, water supplies and other utilities; (3) social infrastructure such as legal, financial and intellectual property rights systems, the degree of deregulation, and governmental institutional framework; and (4) incentive schemes provided by governments such as tax allowances and subsidies for investment and exports. The Flowchart Approach was also verified rigorously by empirical studies such as Tsuji et al. (2006), and Tsuji et al. (2008), for example. These studies also identified concrete factors among those mentioned earlier which actually promoted agglomeration in these regions. Kuchiki and Tsuji (2008) also attempted to develop an alternative to Porter’s framework. They considered that it is difficult to shape practical public policy based on Porter’s framework, since it gives only a partial snapshot of the relationship between industrial agglomeration and local growth. The introductory chapter of Kuchiki and Tsuji (2008) therefore developed a novel and practical model based on the Flowchart Approach that can foster innovation through industrial agglomeration. An advantage of this model

4

Masatsugu Tsuji and Akifumi Kuchiki

is that it can detect the bottleneck effects of agglomeration on innovation, because the authors have already identified the effects of capacity-building and the presence of anchor and related firms on innovation. This model also allows the precise identification of targets for policy implementation or intervention. Although several chapters of Kuchiki and Tsuji (2008) suggest the next research directions of agglomeration, little is known about the endogenous R&D and innovation mechanism. Most recently, Tsuji and Ueki (2008) and Ueki et al. (2008) find the local innovation process due to technical cooperation between industry and local university, R&D institute or local business organization in Indonesia, Thailand and Vietnam. These papers present the endogenous innovation process generated by local alliances with research institutions, the public sector or local companies spread across these developing countries which are collaborating with MNCs in terms of R&D activities. The aim of this book is to uncover some causal relationships between agglomeration and innovation activities in developing economies.

1.2 Research question 1.2.1 Uncovering endogenous R&D and innovation mechanism The research question of this book is: What are the institutional and economic factors, such as government leadership, global competition and comparative advantage, supporting industries that affect promoting industrial clusters including process innovation at the different stages of their formation? The local innovation system requires schemes for clustering and R&D activities. This research question is quite essential to understand clusters and cluster policy. This book also considers the possibility of public intervention and how to foster industrial clusters towards R&D and innovative activities. The Flowchart Approach to the industrial cluster policy can suggest one pathway to simplify the situation from the agglomeration stage to innovation. This book consists of ten chapters, including the introduction and nine chapters to present case studies of industrial agglomeration or innovation processes occurring in particular economies or regions. Underlining this research question, this book addresses following several specific sub-questions related each other. Chapters 2 and 3 describe the sequential relationship from agglomeration to innovation. The case studies of strong initiatives of Malaysian and Singaporean industrial policy would suggest a benchmark for this research. It is clear that this research question is too abstract to understand the local innovation system and to construct policy recommendations based on empirical results. This is why this book offers two specific subquestions to study the local innovation system. Two specific sub-questions emphasize the role of local public policy to stimulate upgrading clusters. First, we focus on the learning linkage between local universities and local business

Introduction 5

organizations as a driver of the endogenous R&D and innovation mechanism, which is referred to as ‘the local innovation system’. The following is the main theme of Part I of this book. Sub-question 1: What is the local innovation system? Can we identify the system in developing as well as developed economies? What are the key factors for forming this system? These questions lead to the following: Since local SMEs can choose their R&D partners, which affects the process of innovation among research institutes (local universities) or large MNCs, to what extent do comparative values of the innovation due to university–industry linkages vs innovation due to outside head-offices generate differences of local alliance activities? Chapter 2 provides the basic framework of the Flowchart Approach and emphasizes localized linkages between university/research institute and business by analysing the case of the automobile industry cluster in Malaysia. Chapter 2 also shows the condition of transition from agglomeration stage to local innovation stage. The next four chapters respond exactly to this question based on the basic framework of Chapter 2. Chapter 3 highlights key challenges and relevant policy implications for promoting university–industry linkages at different stages of industry development by using Singaporean cases. Chapter 4 directly offers empirical evidence on the effect of the proximity of research institutes to SMEs by using case studies of SMEs inside or outside of a cluster in Japan. Chapter 5 attempts to answer this sub-question by studying how the local production system and local alliances between the university–industry linkages drive redevelopment of an ICT (information and communication technology) cluster in Rio de Janeiro, Brazil. Chapter 6 translates the above sub-question to how and what type of information linkage affects process innovation in the Bangalore software industry in India. Second, this book attempts to analyse the agglomeration force as a driver of the endogenous R&D and innovation mechanisms. The following subquestion is not directly related to the university–industry linkage as a source of upgrading. Part II of this book tries to answer following new question. Sub-question 2: How does agglomeration affect business organization and market structure in the stage of upgrading in a cluster? In other words, to what extent does the local government play an important role in the market-oriented innovation process which is driven by agglomeration? This question, left untouched not only by Kuchiki and Tsuji (2005, 2008) but also by the earlier part of this book is focused on here from the viewpoint of the learning linkage between university and business. Chapter 7 studies this by focusing on how organizational renovation differs across developed and developing SMEs located in different clusters in Japan. Chapter 8 examines the effect of local government’s leadership to upgrade

6

Masatsugu Tsuji and Akifumi Kuchiki

firm and market organizations by analysing the micro-structure of the retail market. Chapter 9 also examines labour market pooling effects on building innovation capacity across clusters. The importance of understanding transactions in the retail and labour markets is presented. Although causes and consequences of the coordinating cluster policy are shown in earlier chapters of this book, Chapter 10 focuses on a single firm and its supporting industries to examine agglomeration economies (resource or locationspecific shock to firm) from the inside evidence of innovation decision in a single firm. 1.2.2 Research methodologies This book focuses on identifying factors that encourage agglomeration and innovation and on their effectiveness to regional growth in East Asia and emerging economies such as India and Brazil. In order to investigate the different sources and consequences of innovation in industrial clusters, we utilize different methodologies in such ways as examining public datasets, conducting in-depth interviews and applying econometric analysis by using original survey data. The difficulty of this research lies in postulating the endogenous innovation process; innovation itself is a complex process which encompasses how new tacit ideas are born in specific clusters, and how they end up creating new concrete products, technologies and business models. In fact, not all new technologies are successfully introduced in actual production processes or embodied in new products. We know thousands of examples of these failures, due to all kinds of reasons. One is related to the fact that innovation is a social matter; namely it is related to ways and means of businesses, and not simply related to technology. Typical examples are found in Silicon Valley, Route 128 around Boston and Silicon Highland in Austin, Texas. The rise of these clusters since the early 1990s can be explained not only by the creation of high technology in the fields of IT (information technology), biotechnology, and so on, but also by the creation of new business models, which represent the successful collaboration of technology and business, inspired by enthusiastic entrepreneurs. It is said that the success of Silicon Valley lies in informational networks of technology, funds and human resources which originally existed in the region; and new IT, for example, was just a clue to new business models. Any attempt to postulate the innovation process therefore requires the identification of methodologies which can explain these social phenomena. The innovation process in East Asia and other emerging economies, on the other hand, is different from those in developed economies, since its backgrounds are entirely different. The latter is deficient in conditions for innovation such as technology, funds, human resources with full of entrepreneurships for start-ups as well as engineers. The endogenous innovation process in those

Introduction 7

economies and regions is still a ‘black box’, which we attempt to open in this book.

1.3

Findings

We now briefly introduce these chapters one by one. In Chapter 2, we apply the Flowchart Approach to analyse Malaysia’s automobile cluster policy, investigating whether the industrial cluster policy is successful or not, and proposing a method to prioritize policy measures. According to the Flowchart Approach, the following three policy prescriptions can be recommended: (1) Malaysian firms should establish facilities for exporting compact cars with automatic transmissions; (2) actors in the public, semipublic and private sectors should work together to promote labour skills; and (3) central government should take initiatives in liberalization and deregulation to attract foreign firms into supporting industries. In Chapter 3, we examine the formation of a biomedical sciences (BMS) cluster and an offshore marine engineering cluster in Singapore. By comparing and contrasting these two clusters in the early and more mature stages, we highlight key challenges and relevant policies for promoting industrial clusters at different stages of their formation. Since key processes for cluster development are closely similar for both new and existing clusters, there are common elements in the strategies used in their development. There are, however, distinct differences in the specific role and timing of state commitment depending on the stage of development and nature of a cluster. Chapter 4 attempts to verify the hypothesis that a relationship exists between innovation and industrial clustering formed by regional SMEs and to identify factors which promote SME upgrading and innovation. By comparing the SMEs located inside and outside a cluster, we analyse how industrial clusters and regional research institutions influence the innovation and upgrading of regional SMEs. The following results are obtained: (1) in the most recent five years, location inside an industrial cluster positively influenced the upgrading of SMEs, while before 1998, SMEs located at a distance of 30–60 minutes from a regional research institution were positively influenced; and (2) with regard to innovation, SMEs located over 2 hours from the research institutions tended to achieve fewer innovations. We thus draw the conclusion such that factors influencing upgrading were different at the different stages of agglomerations. In Chapter 5, we present a unique addition to the second phase of the Flowchart Approach applied to a research programme, obtained by studying conditions for the redevelopment of a non-hierarchical cluster in the ICT sector for innovation, with a particular focus on software and services industries in the metropolitan region of Rio de Janeiro, Brazil. We argue that promotion of the territorial agglomeration of high-tech SMEs in local

8

Masatsugu Tsuji and Akifumi Kuchiki

production systems can be sustained by external economies and collective goods that induce them. In Chapter 6, we discuss the information linkage in the process of innovation by analysing the software industry in Bangalore, the largest software cluster in India. Agglomeration and clustering create dense networks that can facilitate face-to-face interactions among individuals and firms. In turn, these networks help them to generate, share and diffuse knowledge, thereby inducing innovation. We examine offshore-based software clusters in India, which mostly engage in software development and services outsourced from foreign clients, and which develop their innovative capabilities without face-to-face interactions with these clients. Using an extensive case study of key firms in Bangalore, this chapter concludes that local innovation systems (LIS) drive innovation in knowledge-based clusters, and that innovation occurs even in the absence of close, face-to-face interactions. Instead, various forms of external linkage serve as vehicles for transferring and diffusing knowledge, and thus for inducing innovation. In Chapter 7, we analyse factors of organizational innovation generated by ICT (information and communication technology) using the survey data of two SME groups, namely developed and developing SMEs in the use of ICT. Since the degree of ICT use and resulting organizational changes differ among SMEs, this chapter constructs a simple index of innovation in ICT use by utilizing AHP (Analytical Hierarchical Process). Based on this index, we attempt to identify factors and policy measures which contribute to innovation by using rigorous econometric analysis, and find that larger SMEs tend have a higher index and top management, who utilize data for management and recognize that the importance of ICT is significantly related to higher use of ICT. In Chapter 8, we analyse the cases of the three largest specialized markets for daily necessities, eyewear and moulds, and find that the local government has played an important role in the development of these specialized markets. As a result, the specialized market has evolved from a typical marketplace into a platform that enables the market mechanism to work more smoothly within the industrial cluster. The experience of these specialized markets shows that markets can take the place of big business in upgrading certain industrial clusters in developing countries. In Chapter 9, we investigate the effects of poaching externality on workplace training using a personnel dataset of different types of industrial clusters such as for food, auto parts, PCs and HDDs (hard disk drives) with different turnover rates in greater Bangkok, Thailand. The empirical analysis suggests that: (1) training incidence is infrequent in industrial clusters with higher turnover rates; and (2) return to training duration is lower in industrial clusters with higher turnover rates. We also show how to extend this approach to estimate the impact of local public policy that fosters industrial agglomeration on product and process innovations through workplace training in

Introduction 9

developing economies. This chapter has a policy implication of the importance of skilled labour supply from local universities to deal with less investment in firm-provided training. Therefore, this implication is a complement of the Flowchart Approach that emphasizes university–industry linkages. In Chapter 10, we examine innovation in an oil and gas industrial cluster in the Campos Basin, Brazil, focusing on its exploration and production segment. We discuss relations between geographic agglomeration and the positive impact on innovative activities by some firms in the Campos Basin cluster, and provide a brief review of policy efforts of local and national actors to maximize benefits accruing beyond the Campos Basin cluster which result from the expansion cycle of the oil and gas industry.

1.4 Summary and policy discussions Let us summarize briefly the main contributions and findings in this book. This book studies internal and external linkages between the production side and the information source of industry upgrading. There is much difference in the degree of dependence on internal and external linkages among clusters. This book finds that the internal linkage which shifts from agglomeration to innovation plays an important role for clusters even in developing countries. The case of Bangalore suggests the emergence of an endogenous R&D process due to vertical specialization and spin-offs. This also emphasizes the importance of internal linkage through the production process, as seen in other chapters. On the other hand, the case of Bangalore provides an exceptional example, since the rapid evolution of Bangalore depends on external linkages through orders, sales, marketing and delivery. By comparing the Bangalore case with others, the large demand from outside the cluster crucially determines the growth of local business and the extent of specialization. Chapters in Part I investigate the extent of the local innovation system through focusing on universities and research institutes as the sources of innovation. Findings in this book lead to policy suggestions that emphasize the role of platform between universities and industries or between industries. Public policy is preferred only in areas such as building and maintaining a platform as a coordinating device between academia and industry. The role of the platform is not restricted only to the university–industry linkage but also to coordinating matching in the retail and labour market. Building a platform for coordinating a firm’s investment behaviour is required to upgrade a cluster. This book also places focus on public policy. Comparison of Silicon Valley-type clusters with Asian-type clusters reveals many differences in the role and efficiency of public policy between the two. The cause and consequences of appropriate public policy depend on the internal structure of an industrial cluster, such as the degree of competition, collaboration, market

10 Masatsugu Tsuji and Akifumi Kuchiki

size or infrastructure. In general, policies for innovation differ between large firms and SMEs. For the former, the typical policy is to construct a so-called ‘National Innovation System’, by combining all R&D organizations including public, private and university research institutions so as to mobilize all forces such as basic applied researches, funding and human resource management in R&D by targeting long-term national objectives. As for the latter, we can identify several examples. Singapore, for example, provides local platforms to facilitate the establishment of alliances between universities and cross-industry organizations. The Japanese Ministry of Economy, Trade and Industry has also been implementing a cluster policy called ‘TAMA’ for SMEs with high technology in the Tama area, a suburb of Tokyo. In these examples, policymakers seek various effective industrial cluster policies such as capacity-building of local SMEs and human capital accumulation in the local labour market to facilitate the ability of the region to innovate. The roles of central and local governments are also important in the context of the clustering policy. A division of roles and collaboration are both required, as already mentioned. With regard to the clustering policy for specific regions, the role of local governments is now much more important than before, since most policy measures to promote agglomeration and regional innovation have been implemented by the central government. Local governments are required to invent new policies to facilitate the regional innovation process.

References Arita, T., M. Fujita, and Y. Kameyama (2006) ‘Effects of Regional Cooperation among Small and Medium Sized Firms on Their Growth in Japanese Industrial Clusters.’ Review of Urban and Regional Development Studies, Vol. 18, No. 3, pp. 209–228. Hashimoto, J. (1997) ‘Nihon-gata Sangyou-Shuseki Saisei no Houkousei,’ in Nihongata Sangyou-Shuseki Saisei no Miraizou, T. Kiyonari and J. Hashimoto Eds, Tokyo (in Japanese), Japan: Nihon Keizai Shinbunsha, pp. 160–198. Kuchiki, A. and M. Tsuji (eds) (2005) Industrial Clusters in Asia: Analysis of Their Competition and Cooperation. Basingstoke: Palgrave Macmillan. —— (eds) (2008) The Flowchart Approach to Industrial Cluster Policy. Basingstoke: Palgrave Macmillan. Markusen, A. (1996) ‘Sticky Places in Slippery Space: A Typology of Industrial Districts.’ Economic Geography, Vol. 72, No. 3, pp. 293–313. Porter, M. E. (1998) The Competitive Advantage of Nations. New York: Free Press. —— (2000) ‘Location, Competition and Economic Development: Local Clusters in a Global Economy.’ Economic Development Quarterly, Vol. 14, No. 1, pp. 15–34. Saxenian, A. (1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Cambridge, MA: Harvard University Press. Tsuji, M., E. Giovannetti, and M. Kagami (eds) (2007) Industrial Agglomeration and New Technologies: A global perspective. Cheltenham: Edward Elgar. Tsuji, M., S. Miyahara, and Y. Ueki (2008) ‘An Empirical Examination of the Flowchart Approach to Industrial Clustering in Greater Bangkok, Thailand,’ in Kuchiki and Tsuji (2008), pp. 194–261.

Introduction 11 Tsuji, M, S. Miyahara, Y. Ueki, and K. Somrote (2006) ‘An Empirical Examination of Factors Promoting Industrial Clustering in Greater Bangkok, Thailand,’ in Proceedings of 10th International Convention of the East Asian Economic Association (CD-ROM), Beijing, China. Tsuji, M and Y. Ueki (2008) ‘Consolidated Multi-country Analysis of Agglomeration,’ in Industrial Agglomeration, Production Networks and FDI Promotion, M. Ariff Ed., Chiba, Japan: Institute of Developing Economies (IDE/JETRO), ch.5, pp. 190–222. Y. Ueki, T. Machikita, and M. Tsuji (2008) ‘Fostering Innovation and Finding Sources of New Technologies: Firm-Level Evidences from Indonesia, Thailand and Viet Nam,’ in Industrial Agglomeration, Production Networks and FDI Promotion, M. Ariff Ed., Chiba, Japan: Institute of Developing Economies (IDE/JETRO), ch.6, pp. 223–289.

This page intentionally left blank

Part I Learning Linkage as Drivers for Creating Cluster and Innovation

This page intentionally left blank

2 The Automobile Industry Cluster in Malaysia Akifumi Kuchiki

2.1

Introduction

Industrial cluster policies have been put into practice in many countries around the world, including in Asia. Our previous studies, such as Kuchiki (2005) and Kuchiki and Tsuji (2005, 2008) postulated a hypothesis regarding the formation of industrial clusters which is referred to as the ‘Flowchart Approach’. This approach explains the formation in such a way that anchor firms establish production bases first; followed by supporting firms establishing facilities near them, which constitutes the core of the industrial cluster at the initial stage, and then in accordance with clustering, more firms agglomerate in those regions and accordingly more information related to transactions, technologies, know-how, etc. are exchanged. This process leads economic activities of industrial clusters to increase more and more, and they come to play important roles such as the economic development of regions as well as a national economy. The Flowchart Approach is thus based on the success of Industrial Parks, Special Economic Zones, or Special Export Zones in East Asia such as Taiwan, Korea and China. The Development of their electronics and automotive industries are good examples. The Flowchart Approach emphasizes factors which attract firms to particular regions, including: (1) domestic demand and natural resources such as raw materials and human resources; (2) physical infrastructure including highways, roads, airports, electricity, water supplies; (3) social infrastructure such as legal, financial and intellectual property rights systems, and the degree of deregulation; and (4) incentive schemes for investment provided by governments. The Flowchart Approach as a principle of constructing policy measures involves setting a proper target, prioritizing the policy measures, and finding actors to implement the policy measures. Let us discuss the difference between the Flowchart Approach and other theories in order to clarify the former. First, Komiya et al. (1988) define the industrial policy as a policy under which the central government intervenes in the existence of market failures due to dynamic inefficiency by 15

16

Akifumi Kuchiki

protecting strategic industries, for example, when they are young from foreign competitors. The Flowchart Approach, on the other hand, emphasizes local governments as well, which play a crucial role in the success or failure of the cluster policy, since it is not a national growth strategy but a regional one. The role of local governments is increasing relative to that of central governments, as the worldwide trend toward decentralization is bringing a shift from industrial policy to industrial cluster policy. Second, Porter (1998) constructed a diamond model, finding that four factors – (1) demand conditions, (2) factor conditions, (3) firm strategy, structure and rivalry, and (4) related and supporting industries – offer sufficient conditions for innovation in industrial clusters. However, it is not easy to satisfy the four conditions at the same time. The aim of the Flowchart Approach is to prioritize the four factors in order, not on a diamond plane but in a line. Markusen (1996) classified industrial districts into three types: Marshallian industrial district, Hub-and-spoke, and Satellite platform. In the case of the Hub-and-spoke type, she found a relationship between anchor firms and their related firms. Markusen (1996), however, neither provided conditions for the formation of industrial clusters nor prioritized factors of the conditions. Kuchiki (2005) proposed the Flowchart Approach of the industrial cluster policy by ordering and prioritizing policy measures in a line for the practical use. Third, in comparison with spatial economic theory, such as Fujita (2008), it presents general equilibrium models, while the Flowchart Approach is a partial equilibrium model, and the above four conditions which determine the success of the policy exogenously determined by other models. In this sense, the Flowchart Approach and spatial economics are complementary. In sum, it can be said that no definite practical method for prescribing the industrial cluster policy has yet been established. This chapter, therefore, makes an attempt to challenge this problem by examining and elaborating further the Flowchart Approach by taking the Malaysian automobile industry as a case study. In East Asia, Thailand continues to grow as the ‘Detroit of Asia’, and Malaysia, on the other hand, plans to establish automobile clusters by implementing an industrial cluster policy, which is referred to as the ‘National Car Project’. Malaysia’s policy, however, has not been as successful as that of Thailand, which offers a good opportunity to discuss whether the industrial cluster policy is effective or not from the viewpoint of the Flowchart Approach. Let us describe briefly the industrial policy aimed to foster the automobile industry in Malaysia. Malaysia’s government started a national car project in 1981. Proton was established as a national car company in 1983. Perodua was established as a second national company in 1993. The Third Industrial Master Plan of Malaysia (2006) referred to automobile clusters, and the Ninth Malaysia Plan of 2006–2010 discussed industrial clusters including automobile clusters in more detail. But Proton faced the difficulty

Automobile Cluster in Malaysia 17

of managing without a joint venture with a company such as Volkswagen in 2006. It is said that Proton will not be competitive in the world without a joint venture with a foreign company. The Malaysian automobile cluster is not so competitive with the other industrial clusters, including those of Thailand in ASEAN. Thus we need to examine what are the conditions of developing the automobile cluster in Malaysia in order for its cluster policy not merely to protect an inefficient automobile industry. In order for the Flowchart Approach to be free from this issue, this chapter will propose the continuous application of the Flowchart Approach in the course of implementing the policy, that is, in addition to using decision-making based on the Flowchart Approach at the planning time, it should be applied at the time that problems occur, so as to solve them. In order to analyse the Malaysian automobile industry, the following methodology is adopted. In step 1, using questionnaires provided by the Japan Bank for International Cooperation and Japan External Trade Organization, we determine which factors along the flowchart process constitute problems of industrial cluster policy. In step 2, we determine policy measures and actors for solving the problems, based on interviews with professionals in the automobile industry of Malaysia. In step 3, we offer prescriptions for the industrial cluster policy by specifying the policy measures and actors along the flowchart. The conclusions obtained by the utilization of the Flowchart Approach which aims to specify actors and prioritize policy prescriptions are summarized briefly in the following three policy measures: first, firms in Malaysia should establish sites for exporting compact cars with automatic transmissions; second, organizations in the public, semi-public and private sectors should endeavour to upgrade skilled labour; and third, the central government should deregulate to attract foreign firms in supporting industries. This chapter is organized as follows: Section 2.2 describes our flowchart model; Section 2.3 explains Malaysia’s industrial policy and automobile industry cluster policy; In Section 2.4, results of survey data and in-depth interviews regarding the competitiveness of the Malaysian automobile industry are presented; Section 2.5 proposes prescriptions for the automobile industry cluster policy from the viewpoint of the Flowchart Approach; Section 2.6 refers to university–industry linkages and national innovation systems; and Section 2.7 concludes the chapter.

2.2

Patterns of the flowchart models

This section proposes the Flowchart Approach for promoting agglomeration and innovation among firms in high-technology and automobile industries. We propose some sufficient conditions for the industrial cluster policy to lead firms to agglomerate in step 1, and innovate in step 2, as is shown in Figure 2.1.

18 Akifumi Kuchiki

(a)

Industrial zone

(b)

Capacity building (I)

Step 1: Agglomeration

(c) (d)

Anchor firm Its related firms Step 2:

(a)

Universities/research institutes

(b)

Capacity building (II)

(c)

Anchor persons including spin-offs

(d)

Innovation (University–industry linkages)

Innovation

National innovation systems

Figure 2.1

Flowchart Approach to industrial cluster policy

Source: Author.

2.2.1 A prototype flowchart model for cluster policy We cannot strictly prove, using an inductive method or deductive method, our hypothesis that the Flowchart Approach will be useful. Our aim is to propose sufficient conditions for the success of an industrial cluster policy. That is, we hope to provide a flowchart that can lead to the successful formation of an industrial cluster if the sufficient conditions listed in the flowchart are satisfied. It should be noted that we can provide illustrations of cases where our hypothesis of the Flowchart Approach holds. We can show a large number of cases, but cannot prove that our hypothesis is a sufficient condition using inductive and deductive methods. Furthermore, our flowchart cannot show that other orderings of factors different from that of our flowchart will not work. Nevertheless, we can show, by increasing the number of cases, that our flowchart may be generally applied to industrial cluster policy in other regions. Our hypothesis is a practical one, since we can form a cluster by following four steps. First, we determine the ingredients of A, B, C, D and E. Second, we select the minimum number of factors from the ingredients found above needed to form a flowchart (see Figure 2.2). Third, we order them along the flowchart (see Figure 2.3). Fourth, we specify actors to proceed at each step of the flowchart if the step goes to ‘No’ rather than to ‘Yes’ (see Figure 2.4). Our flowchart of the automobile industry cluster policy proceeds as follows. First, a local government establishes an industrial zone to attract foreign investors. Second, the government builds the capacity for improving

Example C

Industrial zone

C Yes

A

Capacity building

Actor Local government

A No

E

Anchor firm

Cluster Figure 2.2

Cluster Factors

Source: Author.

Figure 2.3

Cluster policy

Source: Author.

E

Cluster

Figure 2.4

Actors

Source: Author.

20

Akifumi Kuchiki

the business and living conditions for foreign investors. The elements of capacity-building include: (1) constructing physical infrastructure, (2) building institutions, (3) developing human resources, and (4) creating living conditions amenable to foreign investors. Physical infrastructure refers to roads, ports, communications, and so on. Institutional building, which is also crucial for success in inviting foreign investors, includes streamlining investment procedures through one-stop services, deregulation and the introduction of preferential tax systems. Human resources, which are usually an initial condition for foreign investors, include unskilled labour, skilled labour, managers, researchers and professionals. The living environment, for example, includes the provision of hospitals and international schools in order to attract foreign firms. An anchor firm will be ready to invest after this capacity-building has been carried out. 1. Step 1 – Agglomeration: The Flowchart Approach is illustrated in Figure 2.5. First, we ask whether industrial zones have been established. If they have not, we must decide which actors should establish such zones. Once these actors are identified, we return to the main stream of the flowchart.

Industrial zone

No

Yes

Local gov. Return

Central gov.

Semi-gov.

1

Water

2

Electricity

2

Communication

2

Transport

3

Institutions

4

Yes

Figure 2.5

1

2

2

No

1 Return

Flowchart Approach: Step 1: Agglomeration

Source: Author.

Private firms

1

Human resources

Living conditions

NPOs

Automobile Cluster in Malaysia 21

Next we apply the flowchart’s second step, capacity-building, which takes place after the establishment of industrial zones. We examine whether there is an adequate water supply for the industrial zones (see Figure 2.6). We then proceed along the flowchart to examine power supply, communication and transportation. After looking at the physical infrastructure, we examine whether institutions are in place. The central government must institutionalize national tax systems and the local government must institutionalize local tax systems. It is well known that one-stop investment procedures are crucial for success in attracting foreign investors. In the area of human resource development, an abundance of unskilled labour with a high literacy rate is a necessary condition for luring foreign investors whose purpose is to employ cheap labour. On the other hand, an industrial cluster sometimes faces a shortage of skilled labour after industrialization has progressed; universities and on-the-job training centres for innovation are then needed for further development.

Does water supply exist sufficiently?

No

Building

Find players

Assign local govt. priority

2

Assign central govt. priority Assign semi govt. priority Assign NPOs priority Assign private companies priority

Yes

Assign local govt. priority Does electricity supply exist sufficiently?

No

Find players

2 1

Assign central govt. priority Assign semi govt. priority Assign NPOs priority Assign private companies priority

Yes Does communication infrastructure exist sufficiently?

Assign local govt. priority No

Find players

2

Assign central govt. priority Assign semi govt. priority Assign NPOs priority Assign private companies priority

Yes

Assign local govt. priority Does transport infrastructure exist sufficiently?

No

Find players

Assign central govt. priority Assign semi govt. priority Assign NPOs priority Assign private companies priority

Figure 2.6

Flowchart Approach: Step 2: Infrastructure

Source: Author.

3

22

Akifumi Kuchiki

Living conditions are crucial for attracting foreign investors. Researchers from investor companies have incentives to work hard if they can enjoy their lives, so it is important to create satisfactory conditions in areas such as housing, schools, hospitals, and so on. These are the final conditions that must be satisfied to bring in anchor firms. 2. Step 2 – Innovation: Intellectual property rights should be enforced for step 2 of innovation. Preconditions for step 2 are as follows: (1) Related services: finance and insurance, logistics, marketing companies, repair shops, used car shops; (2) Professional and other services: lawyers, restaurants, retail shops, tourism. As is shown in Figure 2.1, the factors of proceeding step 2 of innovation are: (1) universities and research institutes; (2) capacity building of infrastructure, institutional reforms, human resources and living conditions; (3) anchor persons. Then joint actions for step 2 may be as follows: (1) facilitate cluster skill centres; (2) establish collective projects; (3) create business associations; (4) take branding strategy. Linear instruments and interactive approach of policy instruments for innovation may be as follows: Linear instruments: (1) direct R&D aids; (2) transfer of research-based knowledge to firms; (3) financial support; Interactive approach: (1) improving institutions and programmes providing technology transfer services; (2) policy to stimulate networking and business clusters. We illustrated the minimum number of factors of (1) universities and research institutes, (2) capacity-building, and (3) anchor persons to simplify the flowchart of step 2 and prioritize policy measures. Most of Asian countries are located at the entrance of step 2 of innovation and we cannot find a lot of the experiences of step 2 in Asia. Step 2 is still a hypothesis to be further examined. Figure 2.7 summarizes the priorities of each player. Local governments play the main role in establishing industrial zones, supplying electricity, facilitating transport, and reforming institutions. The first priority of local government in Figure 2.7 is to construct industrial zones to accept foreign investors. The second priority during that stage is to supply electricity, facilitate transportation and reform institutions. The main priorities of the central government are to supply electricity and reform institutions. The first priority of semi government in Figure 2.7 is to construct industrial zones. The second priority is to develop human resources. The first priority of NPOs is to improve living conditions. The typical industrial cluster policy was theorized by defining an industrial zone as ‘quasi-public goods’, and it was shown that the policy enhances economic growth under a production function of ‘increasing returns to scale’ of an anchor company. The critical amounts of the production of ‘scale economies’ that are used by related companies to decide whether or not to invest in clusters were also shown (Kuchiki 2008).

Automobile Cluster in Malaysia 23

Local government Priority 1 = establishment of industrial zone Priority 2 = supply of water, electricity & communication Priority 3 = transport Priority 4 = institutional reforms

Central government Priority 1 = electricity supply Priority 2 = institutional reforms

Semi government Priority 1 = establishment of industrial zone Priority 2 = human resource development

NPOs Priority 1 = improvement of living conditions Figure 2.7

Priorities of each player

Source: Author.

Tsuji et al. (2006) pioneered the development of an econometric method giving priority to policy measures. They found support for the prioritization of the Flowchart Approach except for the fact that infrastructure was not found to be a significant factor. 2.2.2

Demand condition

It should be noted that the Flowchart Approach explicitly discusses the demand side of the manufacturing industry which it targets, and there are two issues to be discussed: Case 1: 1. An anchor firm sets up in an export-processing zone and exports its products. In this case, there is little constraint on its demand, since its factory can attain its minimum optimal level of production by exporting

24 Akifumi Kuchiki

products to the world. Here, the logistics of the anchor firm are crucial to attaining the minimum optimal level. 2. Suppliers to the anchor firm can therefore attain their minimum optimal levels of production. The demand for the suppliers is thus ‘demand derived from the anchor firm’. Case 2: 1. The anchor firm sells its products locally. In this case, the size of the local market should be large enough for the anchor firm to attain the minimum optimal level of production. The anchor firm decides to invest in a region only if it concludes that local demand satisfies this condition. 2. The suppliers’ condition is the same as that in Case 1. In advance, we will conclude that Case 2 does not hold in the case of Malaysia, since its domestic market of automobiles is not sufficient. Malaysia has no choice other than to become an exporter, which will be analysed in more detail. 2.2.3 Difference of flowchart models: The automobile industry and the information technology industry This section makes clear the difference of flowchart models between the automobile industry, the information technology (IT) industry and biotechnology industry. The flowchart model for the IT industry is almost the same as that of the biotechnology industry. Knowledge-intensive industries include the IT industry, biotechnology industry and nanotechnology industry. The flowchart model for the knowledge industry is different from that of the automobile industry. It is noted that the knowledge industry requires, as a precondition, the existence of universities. One large difference between the flowcharts of the automobile industry and the knowledge industry is that anchor firms play an important role in implementing the automobile industry cluster policy, while superstars play an important role in implementing the knowledge industry cluster policy. The reason is as follows. While the knowledge industry requires partnership between intellectuals, a car assembler as an anchor firm is essential in the automobile industry cluster policy, since a car is typically composed of more than 20,000 individual components. Superstars are needed for the knowledge industry cluster policy since knowledge is embodied in human resources, so preferential treatment is required in order to lure superstars. Partnerships between universities, large firms, startup firms, multinational firms and laboratories are required for innovations in digital technology. Superstars were crucial for achieving success in creating the partnership and industrial agglomeration in Austin. The leadership of heads of local

Automobile Cluster in Malaysia 25

governments and worldwide superstars in the knowledge industry is key to the success of a cluster policy. Science parks and capacity-building are necessary for a knowledge industry cluster policy. The following three points must be taken into consideration when implementing a knowledge industry cluster policy. First, the regulations should not be too strict. Second, mechanisms should be in place to link patents and innovations. Third, patentees and innovators need to be matched, for which matching funds are effective. It is noted that local governments play an important role in implementing the cluster policy in both the automobile industry and the IT industry. The success of the cluster policy in Austin, Texas in the United States is based on the leadership of the governor of Texas, and the partnership between the state government and municipal government has played an instrumental role in that success.

2.3

Malaysia’s industrial policy and industrial cluster policy

This section explains the situation of Malaysia and its industrial cluster policy related to the automobile industry. Sections 2.3.1 and 2.3.2 describe the histories of Proton and Perodua as industrial policies of national car projects. Sections 2.3.3 and 2.3.4 quote the industrial cluster policy of Malaysia in the Third Industrial Master Plan of 2006 and the Ninth Malaysia Plan 2006–2010. Sections 2.3.5 and 2.3.6 outline transportation costs between countries in ASEAN and the investment conditions of Malaysia. 2.3.1 History of Proton Malaysia’s national car project was carried out as an industrial policy. In 1981, the Malaysian government proposed a joint venture with Mitsubishi, a Japanese automaker, to build a Malaysian car. The cabinet approved the National Car Project in 1982, and Heavy Industries Corporation of Malaysia (HICOM) signed an agreement with Mitsubishi. Proton, a national car company, was established on 7 May 1983, and its factory was built in the HICOM compound. Its first car, named Proton Saga, was launched in 1985, and in 1986 began to be exported to Bangladesh. Accumulated car production reached 50,000 cars in 1987 and 500,000 cars in 1993. The Proton R&D facility opened in 1993, and in 1996 the Proton was being exported to 31 countries. The national car project is a so-called industrial policy, or selective government intervention policy to nurture national firms. However, Proton cancelled its agreement with Mitsubishi in 2004. Proton’s share of sales in the Malaysian market reached about 90 per cent at the highest but fell to 24 per cent in 2005. In 2006, Proton reduced its car prices in Malaysia along with several other car manufacturers as a part

26 Akifumi Kuchiki

of the government policy to lower car prices. In 2007, Proton found itself facing management difficulties without an alliance with a foreign firm. In 1996, Proton City was established as a base for agglomeration by its related suppliers. The Malaysian government set up barriers to investment by foreign firms in order to protect national cars, including Proton, but the policy may have deterred foreign investors in Malaysia and accordingly hindered the agglomeration of foreign firms. 2.3.2 History of Perodua Perodua, a national project, means ‘the second national car’ in the Malay language. Japanese firms contributed to the capitalization of Perodua in the following ways. When POSB (Perusahaan Otomobil Kedua Sdn. Bhd.) of Perodua was established in 1993, the Malaysian government and Japanese firms invested 73 per cent and 27 per cent of the total capital respectively. The POSB invested 100 per cent of the capital in both PSSB (Perodua Sales Sdn. Bhd.) (Perodua’s marketing company) and its vehicle manufacturing company. The capital structure then changed in December 2001, with the Japanese firm acquiring 51 per cent of the capital in the vehicle manufacturing company (or PMSB (Perodua Manufacturing Sdn. Bhd.)/PEMSB(Perodua Engine Manufacturing Sdn. Bhd.)). In summary, the Japanese company owned 47.04 per cent of the total capital of the marketing company and the vehicle manufacturing company, but 51 per cent of the total capital of Perodua. The production processes consist of mainly pressing, painting and assembling. There are 32 Japanese staff members working in important positions in these production processes (November 2006). Production of Perodua cars grew steadily to 102,000 in 2004, recorded 116,000 in 2005, and was expected to be 134,000 in 2006. The production of Protons was 141,000 in 2004 and 139,000 in 2005, but fell to 102,000 in 2006. The market share of Perodua grew from 25 per cent in 2004 to 34 per cent in 2006, while that of Proton, which was more than 80 per cent at one point, gradually fell to about 24 per cent in 2006. The local contents ratio of Kancil and Myvi, both produced by Perodua, is about 80 per cent, while that of Perodua’s Rusa and Kembara models is about 50 per cent. There are 145 domestic suppliers, of which 59 are local suppliers, with a share of 40 per cent. The 19 Japanese suppliers have a share of 13.1 per cent. Once a process of industrial agglomeration as illustrated in Step 1 is completed, the feedback process illustrated in our flowchart in Figure 2.1 can be used to examine whether further industrial agglomeration is possible. The managerial development of Perodua can be summarized as follows: as a national car, Malaysia side holds majority of stocks in the holding company and marketing company, while in manufacturing, Daihatsu holds 51 per cent so as to excise the leadership in production.

Automobile Cluster in Malaysia 27

2.3.3 The Third Industrial Master Plan (IMP3: MITI (Ministry of International Trade and Industry) 2006) of Malaysia As explained in MITI (2006), the Third Industrial Master Plan of Malaysia will provide an overall development framework for the manufacturing sector and detailed sub-sector plans for the period of 2006–2020. Automotive clusters are to be established in Tanjung Malim (Perak), Gurun (Kedah) and Pekan (Pahang), where producers and suppliers of parts and components, and distribution networks have been established. Other areas with some clustering features include Bertam (Pulau Pinang), Serendah (Selangor) and Pego (Melaka). 2.3.4 Ninth Malaysia Plan 2006–2010 The ninth Malaysian Plan is summarized as follows: 1. Chapter 4 of the Prime Minister’s Department (2006) identifies the following anchor companies for automobile clusters: Perusahaan Otomobile Nasional Berhad (PROTON), Petroliam Nasional Berhad (PETRONAS), Tenaga Nasional Berhad (TNB) and a number of MNCs. It lists more than 200 first-tier vendor companies involved in manufacturing and related activities. 2. While the contribution of MNCs to the electric and electronics industry will remain significant, local investment as well as technological capabilities in existing and new electric and electronics activities will be enhanced. These will leverage developments within the cybercities of Bayan Lepas, Pulau Pinang and Kulim Hi-Tech Park in Kedah (Prime Minister’s Department 2006, p. 120). 3. During the Plan period, shared services and outsourcing (SSO) will be positioned as a major new source of growth. In order to further strengthen the SSO cluster, assistance such as access to funds for joint ventures as well as mergers and acquisitions will be made available (Prime Minister’s Department 2006, p. 121).

2.3.5 Transportation costs between countries in ASEAN Perodua’s production system is planned from the perspective of its supply chain management in Asia. Toyota divides its production systems between China and ASEAN in consideration of the logistics under the current situation of Asia’s regional integration. In other words, it has two systems of supply chain management, one for ASEAN and the other for China. The decrease in the tariff rates of ASEAN under the AFTA (ASEAN Free Trade Agreement) affected Malaysia’s management of firms in ASEAN. Daihatsu’s plants in Indonesia and Thailand now supply parts and components to the Daihatsu plant in Malaysia. In particular, the Indonesian plant supplies parts and components to other countries in ASEAN, partly because the plant in Indonesia is large enough to export them.

28

Akifumi Kuchiki

Decisions on whether the plant in Malaysia imports parts and components from other countries are made based on the costs of transportation and the size of the domestic demand in Malaysia. For example, it is costly to package painted car bodies in order to import cars from Thailand to Malaysia. The outward appearance of cars is an important factor in selling them, since consumers’ choices are made partly depending on the perfection of the outside painted appearance so that the painted bodies must be transported from Thailand to Malaysia without being scratched. Trucks can efficiently carry bodies to realize the just-in-time lean production system, but Perodua cannot use trucks since road conditions from Thailand to Malaysia are not good enough to allow the bodies to be transported without becoming scratched. Ships, however, can transport the painted bodies without scratches, and can lower the average costs by guaranteeing that a minimum number of bodies can be carried. 2.3.6

Investment conditions of Malaysia

Next, we look at the situation of suppliers in Malaysia. One of the factors that determines whether foreign suppliers decide to move to Malaysia is the demand from their anchor firm, or the size of its production. The total demand for cars in Malaysia recovered following the Asian Currency Crisis in 1997 and reached 550,000 per year in 2005. This large increase in automobile sales was not anticipated since the population of Malaysia was about 25 million in 2005. One staff member of an automobile company in Malaysia forecasts that car sales will increase by about 100,000 in the next ten years. It is difficult for suppliers in Malaysia to realize a minimum average cost based on the minimum optimal level of production. This is why Japanese suppliers have hesitated to move into Malaysia. The demand condition for attaining the minimum optimal size level is a precondition for starting the flowchart process in Figure 2.1. We conclude that the automobile industry in Malaysia today, with a production level of 500,000 in its domestic market, does not satisfy the minimum demand condition. There are another two reasons why Japanese suppliers find it difficult to move into Malaysia: (1) labour shortage; and (2) instability in institutions such as tax systems. 1. Labour shortage: The Malaysian government has worked, since the early 1990s, to develop the human resources necessary to nurture suppliers in the automobile industry, and the Japanese government has cooperated with these human resource development projects. Malaysian firms are not so positive to acquire technologies compared with China while Chinese firms acquire Japan’s technologies by them. It is said that in the past Japanese and Korean firms went through the same process of acquiring technology by imitating foreign technology that Chinese firms are now experiencing. But Malaysian firms have generally failed to go through this process. Foreign firms in

Automobile Cluster in Malaysia 29

Malaysia, including Japanese firms, complain about the shortage of skilled labour, so human resource development, or capacity building on our flowchart, is needed before taking the next step forward in the cluster policy. The ratio of automation by robots in Perodua’s production process highlights the shortage of skilled labour. The ratio, which in Japan is 99 per cent, is just 9 per cent in Malaysia, according to our interview with an employee of Perodua in October 2006. He told us that few local workers are skilled in maintaining robots, though Perodua provides components such as cylinders to Proton. The reporter told us that human resource development was needed. 2. Instability in institutions: Malaysia’s tax system has changed over time, and is too unstable to allow foreign firms to forecast profits. The current instability of the tax system is considered to be unacceptable to foreign firms, although such instability can be seen in many developing countries such as the value added tax system of China and the import tariff rates of Vietnam. We can use the case of the tariff system reform in Malaysia in 2006 as an example. Under the reform, tariff rates on automobile components were suddenly raised without notice in 2006, and firms cannot calculate the rate of return on their investment in Malaysia and are reluctant to invest in Malaysia. In the long term, the Malaysian government needs to stabilize the tax system.

2.4 Situation of Malaysia among ASEAN countries from the survey data and in-depth interviews 2.4.1 Future expansion of branch activities: JBIC survey The Japan Bank for International Cooperation (JBIC 2007) surveyed 597 firms, as shown in Table 2.1. General machines, electric and electronics, transport machines, automobiles and precision machinery account for 60.8 per cent of the manufacturing industry, compared to 15.3 per cent for the automobile industry. The table shows changes in countries and regions seen as the most promising place to do business in the three-year medium term. The ratio of votes for Thailand to the total was 28 per cent in 2002, 29 per cent in 2006 and nearly the same figure from 2002 to 2006. The ratio of votes for Vietnam to the total increased from 15 per cent in 2002 to 33 per cent in 2006. Indonesia’s share decreased from 15 per cent in 2002 to 8 per cent in 2006. Similarly, that of Malaysia decreased from 8 per cent in 2002 to 5 per cent in 2006. This section explains why Malaysia’s ratio decreased, and why Thailand’s did not. Figure 2.8 shows that the number of firms in the automobile industry planning to strengthen and enlarge their branch activities in Thailand, Indonesia, Vietnam and Malaysia is 55, 24, 15 and 10, respectively. Malaysia has the smallest number among them.

30 Akifumi Kuchiki Table 2.1 Changes in countries and regions for promising businesses in the medium term (three years) (Unit: %) 2002 (Number of Ranking replies: 418) 1 2 3 4 5 6 7 8 9 10 11

China Thailand USA Indonesia Vietnam India Korea Taiwan Malaysia Brazil Singapore

89 28 26 15 15 13 8 8 8 5 4

2003 (Number of replies: 490) China Thailand USA Vietnam India Indonesia Korea Taiwan Malaysia Russia Singapore

2004 (Number of replies: 497)

93 29 22 18 14 13 9 7 6 5 5

China Thailand India Vietnam USA Russia Indonesia Korea Taiwan Malaysia Singapore

91 30 24 22 20 10 10 9 8 6 3

2005 (Number of replies: 483) China India Thailand Vietnam USA Russia Korea Indonesia Brazil Taiwan Malaysia

82 36 31 27 20 13 11 9 7 7 5

2006 (Number of replies: 484) China India Vietnam Thailand USA Russia Brazil Korea Indonesia Taiwan Malaysia

77 47 33 29 21 20 9 9 8 6 5

Note: The entries in bold refer to Thailand and Malaysia. Source: JBIC (2007) Survey Report on Overseas Business Operations by Japanese Manufacturing Companies. Tokyo: JBIC.

76

80 70 55

50 40

29

24

30 20

10 2

3

3 Hong Kong

10

Singapore

9

Taiwan

No. of replies

60

15 5

Figure 2.8

Vietnam

India

China

Philippines

Malaysia

Indonesia

Thailand

Korea

0

Number of firms to strengthen and enlarge their branches

Source: The same source as that of Table 2.1.

2.4.2

Countries to invest: JETRO survey

The Japan External Trade Organization (JETRO 2006) surveyed 966 firms. Of these firms, as shown in Table 2.2, 44.2 per cent were in general machines, electronic and electronics appliances, electric and electronics parts, transport equipment, and transport equipment, whereas 17.2 per cent were in transport equipment and parts. Of these firms, 937 answered a question on what they considered to be suitable production

Table 2.2 Suitable production sites in the medium and long term (Unit: number of firms) Countries where firms are located

Countries and regions selected for suitable production sites

Country No. of valid (ASEAN/India) replies Thailand Malaysia Singapore Indonesia

Others (Hong Kong/ Philippines Vietnam India China Taiwan)

Thailand Malaysia Singapore Indonesia Philippines Vietnam India

199 169 95 149 180 84 61

124 28 16 21 30 9 9

0 68 5 2 3 1 0

0 0 23 2 0 0 1

4 2 8 53 3 0 3

1 1 0 0 54 0 0

30 30 20 21 50 48 0

20 10 5 18 10 9 38

16 18 14 20 19 7 6

4 12 4 12 11 10 4

Total note

937

237

79

26

73

56

199

110

100

57

Note: (1) Total is the summation of the number of firms that select countries and regions as suitable production sites. (2) The entities in bold refer to the numbers crucial to comparison between Thailand and Malaysia. Source: Japan External Trade Organization, Management Conditions of Japanese Firms in the Manufacturing Industries in Asia, March 2006.

32 Akifumi Kuchiki

sites in the medium and long term. The number of respondents citing Thailand, Vietnam, India, China and Malaysia was 237, 199, 110, 100 and 79, respectively. From these surveys by JBIC and JETRO, we conclude that Japanese firms are expected to invest in Thailand, Vietnam or India, rather than Malaysia. This section elucidates the differences between Malaysia and Thailand in terms of the factors for agglomeration and clustering. Fourin (2009), which illustrates the domestic sales figures of automobiles in the main ASEAN countries, shows that those of Malaysia, Indonesia and Thailand are almost the same at 487,000, 483,000 and 626,000, respectively. JAMA (2005), which gives the automobile production figures of the main ASEAN countries, shows that the production in Malaysia in 2004 was 480,000 while that of Thailand was twice as large, with 920,000. Fourin (2006) shows the difference between the production figures and the domestic sales figures. For Thailand, this figure was 460,000 in 2005, since firms had sites for exporting products. Malaysia cannot have sites for exporting products partly due to the shortage of capacity of the supporting industries. Table 2.3 clarifies the problems in terms of the localization of employees, products and technology. Problems in localizing employees include a shortage of local managers employable at reasonable wages and top managers to promote localization. Both Malaysia and Thailand have a similar shortage of local managers and top managers. Problems in localizing products faced by all the countries shown in Table 2.3 are a shortage of technology at local firms. Problems in localizing technology faced by all countries shown in Table 2.3 include a shortage of local skilled labour at reasonable wages. In sum, the common problems faced by all of the countries are a shortage of local managers, top managers and local skilled labour. Table 2.3 suggests and supports the result that the problems of Thailand appear to be more serious than those of Malaysia. Thus, the shortage of human resources alone is not sufficient to explain the difference of export capacity between the two countries. Table 2.4 shows reasons why certain countries and regions are seen as promising. All ten countries and regions shown in Table 2.4 give high priority to promising local markets. A large local market is a precondition when multinational corporations are deciding on a location. It is expected that in the future, the markets of Vietnam and Indonesia will become larger than that of Thailand. All ten countries and regions shown in Table 2.4 give a high priority to the existence of abundant labour. Table 2.5 shows an index for evaluating the investment environment of ASEAN and India in comparison with China. The index is calculated as follows: A 5 (Number of firms who replied ‘superior’) 2 (Number of firms who replied ‘inferior’),

Table 2.3

Problems in localization of employees, products and technology (Unit: %)

Problems in localizing employees (Number of replies) Shortage of local managers employable at reasonable wages Shortage of top managers to promote localization Difficulty of smooth communication Prevention of outflow of secret information to other firms (including job hopping) Difficulty of making a manual for work processes Construction of staff evaluation system suitable to local employees Others No problem Problems in localizing products (Number of replies) Shortage of technology of local firms Prevention of outflow of secret information to other firms (including the outflow of blueprints) Shortage of competitiveness in costs of Japanese and foreign firms Shortage of competitiveness in costs with local firms Lack of local supporting industries Shortage of technology of Japanese and foreign firms Others No problem Source: The same source as that of Table 2.1.

China (462)

Thailand (265)

Malaysia (145)

Indonesia (152)

Vietnam (93)

India (77)

50.2

45.3

33.1

48.0

44.1

27.3

39.4 37.9 34.0

37.7 27.9 18.9

36.6 18.6 17.2

52.0 28.9 19.1

39.8 32.3 16.1

28.6 26.0 23.4

16.2 16.7

14.0 9.8

12.4 9.7

12.5 12.5

19.4 12.9

14.3 18.2

3.5 8.0

3.8 18.5

8.3 21.4

3.9 9.9

5.4 14.0

14.3 14.3

China (400)

Thailand (218)

Malaysia (116)

Indonesia (136)

Vietnam (78)

India (61)

55.3 35.0

45.0 12.8

32.8 12.9

51.5 12.5

38.5 15.4

41.0 18.0

17.0

12.4

20.7

15.4

17.9

14.8

14.8

15.6

11.2

17.6

16.7

13.1

12.3 9.5

6.9 9.2

11.2 3.4

16.9 11.0

34.6 12.8

24.6 4.9

7.5 14.3

6.0 33.0

5.2 34.5

4.4 20.6

3.8 15.4

13.1 14.8

Table 2.4

Reasons for promising countries and regions (Unit: %)

2006 Survey (Number of replies) Risk averse to other countries Sites to export to Japan Sites to export to third countries Present large size of local markets Growth potential of local markets Profitability of local markets Political and social stability Developed infrastructure Developed logistics services Preferential tax system Stable policies Excellent human resources Cheap labor Cheap parts and materials Supply sites to assembly companies Industrial agglomeration Sites for product development

1. China 2. India 3. Vietnam 4. Thailand 5. USA 6. Russia 7. Brazil 8. Korea 9. Indonesia 10. Taiwan (362) (223) (154) (133) (101) (94) (44) (41) (37) (26) 1.9 15.2 19.3

10.8 2.2 9.4

36.4 11.0 18.2

21.1 12.8 28.6

1.0 2 1.0

4.3 1.1 2

9.1 4.5 11.4

2.4 2 7.3

5.4 18.9 27.0

2 3.8 7.7

24.9

11.7

5.2

24.1

70.3

14.9

15.9

41.5

27.0

50.0

82.3

83.0

46.8

42.1

44.6

93.6

81.8

73.2

59.5

69.2

7.2 1.4 5.8 3.0 13.5 1.4 16.6 57.2 23.5 27.3

4.0 5.8 1.8 0.4 5.4 1.3 35.0 44.4 9.0 21.1

3.9 15.6 3.9 1.9 16.9 8.4 35.1 71.4 5.8 22.7

10.5 24.8 27.8 6.8 24.1 16.5 17.3 45.9 9.0 36.8

21.8 37.6 42.6 24.8 2.0 5.0 15.8 2.0 4.0 18.8

8.5 3.2 4.3 2 4.3 1.1 5.3 17.0 3.2 16.0

13.6 6.8 4.5 2 4.5 2.3 6.8 22.7 6.8 18.2

17.1 7.3 17.1 9.8 12.2 4.9 9.8 7.3 2.4 12.2

13.5 2.7 8.1 8.1 2 2.7 8.1 54.1 16.2 18.9

11.5 3.8 15.4 11.5 19.2 3.8 11.5 15.4 11.5 15.4

16.6 4.4

6.3 2.2

4.5 1.3

30.1 5.3

19.8 12.9

2.1 2

4.5 2

12.2 2.4

10.8 2

11.5 2

Note: The entries in bold refer to crucial factors of the investment environment. Source: Outlook for Japanese Foreign Direct Investment, 2007 JBIC (Japan Bank for International Cooperation) Institute, May 2007.

Table 2.5

Evaluation index of investment environment of ASEAN and India in comparison with China (Unit: %) Evaluation items (2006)

Pre-condition

Agglomeration Note:

∆

Infrastructure Tax system Transparency of laws related to investment Little risk of volatility of currency exchange rates Import procedures Protection of intellectual property rights Ease of personnel labour management Level of research and skilled labour Political and social stability Communication capability of employees Level of development of supporting industries

Thailand

Malaysia

Singapore

64.5 50.4 68.9

67.0 61.7 66.1

95.6 97.1 92.6

∆

13.1

30.2

52.2

∆

42.0 34.2

64.2 38.5

95.6 94.1

∆ ∆

52.0

21.2

85.1

9.5

75.0

∆

65.6

90.8

84.8

95.7

∆

34.6

52.7

88.2

27.5

∆

6.6

22.1

∆

7.4

∆

means negative values.

Source: JETRO, Current Management of Japanese Manufacturing Industries in Asia, March 2006.

Indonesia 59.8 35.2 ∆ 17.2

India

74.6 7.0 6.9

∆

68.1

28.1

∆

14.4 12.1

7.0 6.9

∆

∆

∆

∆

∆

13.2

40.5 39.5 0.0

20.7

33.3

22.7

73.8

50.0

7.4

20.3

71.8

∆

4.4

77.5 13.2 22.5

∆

48.3

∆

∆

Vietnam

71.1

∆

∆

85.2

∆

31.6

36 Akifumi Kuchiki

B 5 (Number of firms who replied ‘superior’) 1 (Number of firms who replied ‘inferior’), and Evaluation Index 5 (A/B) 3 100. The indices for Malaysia and Thailand are very similar, except that the level of development of supporting industries of Malaysia (2 6.6) is much worse than that of Thailand (1 27.5). Table 2.3 also shows that, regarding the problems in localizing products, the lack of local supporting industries of Thailand is 6.9 per cent while that of Malaysia is 11.2 per cent. 2.4.3 Summaries of surveys By examining the questionnaire data from JBIC and JETRO, the following two facts of Malaysia’s automobile industry cluster can be extracted. First, Malaysia has a shortage of unskilled labour for attracting foreign investors. The shortage of skilled labour is a crucial problem in view of the creation of further industrial agglomeration in Malaysia. Malaysia has introduced foreign workers into the palm oil industry and the construction industry, from Indonesia, Bangladesh and other countries, but it is required more, and this relates to the problem of social stability. Second, Malaysia has a lack of developed supporting industries in comparison with Thailand, since one of the factors that prevents foreign investors from investing in Malaysia is its institutional instability, which will be discussed fully in the context of the application of the Flowchart Approach. 2.4.4

In-depth interviews on Malaysia’s automobile industry

In order to derive prescriptions for Malaysian automobile industry by applying the Flowchart Approach, ten professionals related to the Malaysian automobile industry took part in in-depth interviews. The questionnaire used is summarized as follows: 1. Is the size of Malaysia’s automobile market large enough? 2. Can Malaysia host export sites? 3. Will foreign automobile firms invest in Malaysia under the present regulations? 4. Will foreign automobile firms invest in Malaysia despite the shortage of unskilled labour? 5. Can Malaysia catch up with Thailand in its supporting industries? 6. Are the controls over foreign capital inflows of suppliers to the national cars sufficiently deregulated? We found the following six results, as summarized in Table 2.6. (The ten respondents included one staff member of a national car project, two professors at the University of Malaya, one staff member of a national car supplier, two instructors for small and medium enterprises, three members of staff

Automobile Cluster in Malaysia 37 Table 2.6 Questionnaire survey on application of Flowchart Approach to industrial cluster policy (May 2007)

1. Is the size of the market of Malaysia large? 2. Can Malaysia be an export site? 3. Can foreign auto firms invest in Malaysia under the present regulations? 4. Can foreign auto firms invest in Malaysia under the shortage of unskilled labour? 5. Can Malaysia catch up Thailand in supporting industries? 6. Are the controls over with foreign capital inflow of suppliers to the national cars deregulated?

1

2

3

4

5

6

7

8

9

10

%

3

3

3




3

3

3

3

3

3

10

3




3







3




3




3

50

3




3




3

3




3

3

3

30




3






















3

80

3




3

3

3

3




3

3




30

3




3

3

3

3




3




3

30

Source: Author.

of Japanese semi-government organizations, and one Japanese government officer in Malaysia). Size of the market Regarding the size of Malaysia’s domestic car market, Malaysians cannot be expected to buy more cars since the size of the population is small at about 25 million in 2007, per capita income is more than US$5,000, and each family owns one car on average. The total domestic demand for automobiles might not increase from the current level of approximately 500,000 per year. Only two of the ten respondents stated that the demand was not small.

●

Export possibility Regarding the question of whether Malaysia can host export sites, six of the ten answered yes. A division of labour between Thailand, Indonesia and Malaysia is needed for Malaysia to become a host to export sites. Thailand and Indonesia specialize in pick-up trucks and multi-purpose vehicles respectively. Malaysia could specialize in compact cars with automatic transmission, but automatic cars are still not very popular in Asia. However, Malaysia must compete with India, which is expected to increase

●

38

Akifumi Kuchiki

its production of compact cars, and China, which has excess production capacity for compact cars. Foreign investment and regulation Regarding the question of whether foreign automobile firms will invest in Malaysia under the present regulations protecting the national cars such as Proton and Perodua, three of the ten answered yes. Most of the respondents judged that deregulation and liberalization would be required to attract foreign investors. Malaysia should attempt to promote the influx of foreign direct investment through deregulation.

●

Foreign investment and unskilled labour Regarding the question of whether foreign automobile firms will invest in Malaysia despite the shortage of unskilled labour, eight of the ten answered yes. Malaysia has introduced unskilled labour from foreign countries for the past 20 years. The foreign countries were mainly Indonesia and Bangladesh in the 1990s, and Vietnam, Myanmar and Nepal in 2007. Taiwan is similar to Malaysia in introducing unskilled labour from foreign countries. One of the respondents answered that the social cost had increased since the share of foreigners in the total population exceeded 20 per cent in 2007.

●

Supporting industries On the question of whether Malaysia can catch up with Thailand in the supporting industries, three of the ten answered yes. Thailand has actively introduced foreign direct investment, while Malaysia has tried to protect and foster its domestic firms. For this reason, there is a large difference in the development of the supporting industries between the two countries. However, the development of Thailand’s supporting industries has relied on foreign capital, while Malaysia has tried to foster domestic firms. Malaysia is slowly nurturing local suppliers step by step. This question makes clear that it is suggested that Malaysia introduce foreign firms in the supporting industry instead of fostering domestic firms. Some of the respondents recommend that the Malaysian government deregulates the laws to attract foreign firms as the Thai government does. The deregulation is one of policy measures of Malaysia to solve the problem of shortage in the supporting industry.

●

Deregulation Regarding the question of whether the controls over inflows of foreign capital into suppliers of the national cars are sufficiently deregulated, three of the ten answered yes. Some respondents stated that, generally speaking, suppliers of parts, who provide their products to Proton and Perodua, cannot meet world quality standards. Some local suppliers related to the national cars provide their products to Japanese anchor firms in Malaysia, and may be competing with other suppliers in the world.

●

Automobile Cluster in Malaysia 39

2.4.5

Summary of in-depth interviews

Table 2.6 shows the following three results from the questionnaire: first, firms in Malaysia should establish sites for exports; second, skilled labour will not be available in Malaysia; third, in order to attract foreign firms in the supporting industries, deregulation should be required. Some of the respondents of the interviews suggested the following three policy measures. First, firms in Malaysia should establish sites for exporting compact cars with automatic transmissions. Second, actors or players should endeavour to upgrade skilled labour. Third, the government should deregulate rules in order to attract foreign firms in the supporting industries. The prescriptions for the automobile industry cluster policy suggested by the respondents are as follows: (1) Malaysia should aim to become sites for exporting automatic compact cars while firms in Indonesia and Thailand are engaged in exporting pick-up trucks and multi-purpose vehicles, respectively; (2) organizations in the quasi-public sector and private firms as actors should upgrade labour skills rather than nurture unskilled labour; and (3) the central government must further deregulate the national car policy to attract foreign direct investment into the supporting industries in Malaysia (see Figure 2.9). There is no doubt that the above three issues are important to the further development of the Malaysian automobile industry, but a real problem is how to do it. We will then discuss this in the next section by applying the Flowchart Approach.

Capacity building

Anchor firm

Feedback (Combination of the electronics industry and automobile industry) Figure 2.9

Related firms

Partnership with foreign firms

Go to step 2. innovation

Prescriptions for the automobile industry cluster

Source: Author.

Local firms

40 Akifumi Kuchiki

2.5 Prescriptions for Malaysia’s automobile industry from the Flowchart Approach The Flowchart Approach specifies actors or players and prioritizes policy measures, and here we examine how to revitalize the Malaysian automobile industry by utilizing the Flowchart Approach. 2.5.1 Prioritization of targets The Malaysian National Car project has two objectives; (1) fostering the automobile industry and (2) enhancing local Malayan firms and businesses in the automobile and related parts industry. The former is the same as a target of the industrial and cluster policy. Since the automobile industry is highly capital- and technology-intensive, by nurturing the automobile industry the level of technology and labour quality of whole national economy will enhance together due to the external effect, which promotes international competitiveness not only of the automobile industry but also of other industries. Successful examples of these policies are found in those of Japanese, Korean and recent Chinese automobile industries. Malaysia, on the other hand, has another target of (2) above, that is, it strongly assigned a political role of enhancing the economic and political position of bumiputra (native Malays). As motioned already, national cars, namely Proton and Perodua, were heavily protected by low excise tax which created large price discrepancies between non-national cars, high customs to foreign automobiles, local contents policy of automotive parts and various non-tariff barriers. These protections, on the other hand, ended up with inefficient production and accordingly weak international competitiveness of national cars due to lack of competition pressure. The Flowchart Approach prioritizes targets regarding clustering or other particular policy objectives basically according to economic criteria, that is, efficiency and the cost–benefit basis. In retrospect, the Malaysian government should have placed the first priority on the development of the automobile industry per se, and in accordance with its development, other policy measures could have implemented to foster local firms related to supporting industries and native Malayan entrepreneurs. Nowadays, because of AFTA, the automobile industries in ASEAN have been engaging in restructuring their business strategies by aiming at not only ASEAN but also the global market. The Malaysian automobile industry is not independent of these trends, and it is required to select targets from these standpoints. These are policies recommended by the Flowchart Approach. 2.5.2 Selection of anchor firm After the section of a proper target, the Flowchart Approach deals with the selection of anchor firms and related firms. This process is shown in

Automobile Cluster in Malaysia 41

Figure 2.10 for Malaysia’s automobile industry cluster policy. According to the Flowchart Approach, host countries prepare with conditions such as capacity-building to invite MNCs, for example, and foreign firms make a decision over establishing production bases there. In the Malaysian automobile case, the anchor firm was already decided by the government, and foreign firms are positioned as partners to local anchor firms. The partner of Proton was Mitsubishi Motor Corporation, while that of Perodua was Daihatsu. In addition, according to the Malaysian policy mentioned earlier, the balance of power leans toward Malaysian firms, and this spoiled the leadership of foreign partners which owned technological and managerial advantages. In other words, host countries cannot fully make use of their advantages to promote their economic development. The ownership of Proton had changed over time: Heavy Industries Corporation of Malaysia (HICOM) owned Proton when it was established. In 1995, DRB (Diversified Resources Bhd.) merged with HICOM and BDRHICOM (Heavy Industries Corporation of Malaysia Bhd.) was the second owner. In 2000, Proton was bought by Petronas which is a national oil and gas company. These imply that Proton is not proper as an anchor firm, since its business basis was not strong enough. Since 2000, Japanese automobile assemblers such as Toyota, Honda and Nissan increased their presence in Malaysia. They are, however, engaging Stage I. Agglomeration

Domestic demand: 500,000 Actors Capacity building: deregulation of National car policy Human resource development: skilled labor

Central government

Private firms

Quasi-gov.

Agglomeration of suppliers I Feedback process

Export demand: compact automatic cars

Private firms

Capacity building Anchor firms Agglomeration of suppliers II

Figure 2.10

Prescriptions for automobile industry clustering in Malaysia

Source: Author.

42

Akifumi Kuchiki

currently in the global strategy regarding where to assemble, where to manufacture automotive parts, where to sell, and where to conduct R&D. Their interest and objectives in manufacturing in Malaysia is different from in the past. 2.5.3 Size of the market: Demand condition Size of the market matters to the automobile industry, since it is related to scale economies, that is, the larger the amount of automobile production becomes, the more the average cost reduces. The amount of production is thus the source of competitiveness. The rationale behind the Malaysian national car project lay in concentrating automobile production into Proton, which realized the scale economies. In this respect, the policy was right. The size of the Malaysian domestic market is clearly small as shown in replies of professionals to the questionnaire of Table 2.6. Malaysia used to have the largest automobile market in ASEAN from 1997 to 2002 (see Table 2.3) in comparison with its population. Regarding the amount of production, it is smaller than that of Thailand, and in 2004 Malaysia produced less than half of Thailand (see Table 2.4). The gaps between the amount of production and domestic sales are exports. As mentioned earlier, due to the small scale of production, the Malaysian automobile industry did not have international competitiveness. Considering the size of the domestic market, the industry policy fostering the automobile industry should have aimed the international market at the beginning. This problem is related to a previous issue of the selection of anchor firms. If we apply the Flowchart Approach to selecting anchor firms at this moment, when the automobile industry is in the midst of the global restructuring, firms with global networking and competitiveness should be chosen as anchor firms, and the role of the Malaysian position is determined in their framework of global competition. It is interesting to compare with Tata Motors in India, whose international partner is Suzuki Motor Corporation which is quite similar to Daihatsu; the recent development of Tata is supported by the huge Indian domestic market. The global market will provide the same chance to Proton, only if it can compete in the global arena. 2.5.4 Reforms suggested by the Flowchart Approach Here let us apply the Flowchart Approach to the Malaysian automobile industry and examine policies for it to be competitive in the global market. In order to invite anchor firms, namely automobile assemblers and related firms, to Malaysia, it should prepare conditions as the Flowchart Approach requires, which is shown in the beginning of this chapter, which are: (1) domestic demand and natural resources such as raw materials and human resources; (2) physical infrastructure including highways, roads, airports, electricity, water supplies; (3) social infrastructure such as legal, financial and intellectual

Automobile Cluster in Malaysia 43

property rights systems, and the degree of deregulation; and (4) incentive schemes for investment provided by governments. Some of these factors are already satisfied, and some are not. In order to focus on particular factors in more detail, let us examine Table 2.5 which indicates evaluation index of investment environment in Asia. Factor with low index are as follows: (1) level of research and skilled labour; (2) supporting industries; (3) ease of personnel labour management; (4) protection of intellectual property rights; and (5) volatility of foreign exchange rates. Moreover, Table 2.6 indicates the issues raised by interviews with professionals, and the following questions were rated low: 3. Will foreign automobile firms invest in Malaysia under the present regulations? and 5. Can Malaysia catch up with Thailand in its supporting industries? From the above results of two investigations, it is necessary to reconstruct capacity building such as (1) skilled labour; (2) social infrastructure such as legal, and intellectual property rights systems; and (3) deregulation related to assemblers as well as supporting industries, as the Flowchart Approach suggests inviting foreign firms. Reconsideration of the bumiputra policy should be required, since it contradicts the global trend of competition. A good example is shown in transplants of Japanese automobile assemblers in the US. They established assembling plants in the southern states where the influence of UAW (United Auto Workers) was weak, instead of traditional automobile clusters in the northern states. They chose the merit of being free from labour unions, instead of economic benefits located in the centre of automotive clusters. The national car project originally aimed to foster the automobile industry, as mentioned earlier, and as such the industry policy, however, tends to have negative aspects to protect a particular industry from foreign competitors and maintain it in less competitive situations. In order for the Malaysian automobile industry to be competitive, deregulations are necessary to open the market to foreign firms by abolishing heavy excise tax not only on imported cars but also automotive parts. Regarding support to automobile industries, due to local contents, 70 per cent of Proton parts are supplied by domestic parts suppliers, but their technology and quality depend on those of Japanese and other foreign firms. According to the Flowchart Approach, automobile assemblers with various superior parts suppliers should be selected as anchor firms. 2.5.5 Summary of policy prescriptions by the Flowchart Approach Let us summarize the above discussion on policy prescriptions suggested by the Flowchart Approach, which is indicated in Figure 2.10. First, anchor firms or their partners should be reconsidered. Automotive assemblers with strong competitiveness and strong related industries are to be selected. Second, in order to invite those firms, organizations in both the

44

Akifumi Kuchiki

quasi-public sector and private firms must act as actors to engage further in human recourse development. Third, the central government must commit to and enforce the Economic Partnership Agreement (EPA) and Free Trade Agreement (FTA) in ASEAN in order for the Malaysian industry to enhance competitiveness by opening its market. From these, the Malaysian industry can access cutting-edge technology and know-how through foreign direct investment, for example.

2.6 University–industry linkages and national innovation systems This book provides a practical framework to explain the formation of agglomeration and the endogenous innovation process of upgrading industrial clusters to the higher R&D. The flowchart approach to industrial cluster policy consists of step 1 of agglomeration and step 2 of innovation. Figure 2.11 shows the flowchart of step 2 of innovation by using the components of university–industry linkages (UILs) and the country’s national innovation system (NIS). UILs and NIS played key roles in explaining innovation in Hershberg et al. (2007) and Brimble and Doner (2007). Hershberg et al. (2007) surveyed the substantial literature on agglomeration economies and found that, in Asia as well as Western Europe, there were relatively few instances of university– industry linkages (UILs).

Actors Step 2 Innovation

Prioritization University Yes Science park Yes Capacity building

Return No

Spin-offs

2. NIS,RIS,MIS

Education

Anchor person Industry research institutes

o

o

o

o

Return

Return

o o

Agglomeration

o

1. UILs linkages Output = innovation

1.Government 2.University 3.Industry Central Local

U = University/research institutes pre-condition No

o

(note)

2. NIS,RIS,MIS National innovation system Regional innovation system Municipal innovation system

3. Clustering

Figure 2.11

Flowchart of innovation by university, industry and cluster

Note: Innovation = f (UILs, NIS: Cluster) where UILs denotes university–industry linkages, NIS denotes national innovation system. Source: Author.

o

Automobile Cluster in Malaysia 45

Brimble and Doner (2007) found little UILs and weak NIS in the case of Thailand in the following. Thai industry has historically shown little interest in innovation. With some interesting exceptions in particular sectors and organizations, there are few university–industry linkages in place with clear benefits to the public and private sectors. This reflects and contributes to what has been a relatively ‘weak and fragmented’ national innovation system). Chen and Kenney (2007) pointed out that high-tech zones in China were built in close proximity to universities and public research institutes with the goal of promoting UILs. That is, the existence of universities is a precondition of establishing high-tech zones to promote UILs. Wu (2007) analysed UILs in the context of national innovation systems and found that UILs are active exceptionally in Beijing. Three key institutional actors are industry, research organizations (universities/public research institutes: URIs) and government. UILs are built through two broad categories of mechanism. The first is technology transfer through licensing and other arrangements such as consulting, joint or contract R&D, and technical services. The second mechanism is through university enterprises. The unique feature of the Chinese NIS is the URI-owned enterprises such as Lenovo and Founder at the Zhongguancun hi-tech park of Beijing. The Beijing Universities of Peking and Tsinghua have generated spin-offs and established science parks to commercialize their research and technology. The spin-offs have established the agglomeration of firms. The Beijing universities have developed close relationships with industry (UILs) through joint projects, professional consulting and training. Kuchiki (2007) derived the flowchart of step 2 of innovation from the hi-tech zone of the Zhongguancun area of Beijing. However, the cases in Beijing are exceptional and little UILs in China. Wu (2007) said that the local impact of university-based innovation and entrepreneurship should not be overstated. In 2001, only about 40 per cent of university enterprises were involved in science and technology. Their sales revenue made up a mere 2.3 per cent of all high-tech enterprises nationwide. Concerning Malaysia, its Ninth Plan intends to raise the capacity for knowledge and innovation. The efforts will be intensified to develop the country’s human capital in order to drive the transformation to a knowledgebased economy. Programmes and projects will be undertaken to deliver the National Mission’s priorities of improving the education system, increasing innovation and ensuring holistic human capital development. The National Advisory Council on Education and Training will be established to guide policies on strengthening the National Innovation System (NIS). To further encourage innovation, technology transfer and commercialization, the existing intellectual property (IP) framework will be strengthened to enhance IP support facilities and to shorten the IP approval process. The intellectual property right is a precondition for the Malaysian economy to go from step 1 of agglomeration to step 2 of innovation.

46

Akifumi Kuchiki

Malaysia’s automobile industry cluster has not much experienced UILs together with its NIS. Based on Kuchiki (2007), the flowchart of Figure 2.11 summarizes the analyses of the papers introduced in this section. The other chapters of this book will further analyse step 2 of innovation instead of this chapter. Hershberg et al. (2007) stated that, among Asian economies, Singapore is pursuing the US model of the UIL most closely. Poh-Kam Wong analysed the biology industry cluster in Singapore in Chapter 5. We applied the Flowchart Approach to this case, as is shown in Figure 2.12. Figure 2.11 is a prototype model of the Flowchart Approach while Figure 2.12 is a variation of the prototype model. Further analyses on step 2 of innovation are undertaken in the other chapters. The roles of parks, anchor persons and capacity building Biopolis The government organizations

A*STAR EDB BSG BOC

MNCs Outsource

Industrial park

the bio committees NIS (National Innovation System)

GSK Merlion Schering-Plough LSB etc.

Anchor persons etc. Sidney Brenner Yoshiaki Ito etc. Alliances with universities Johns Hopkins MIT etc. with NSU Spin-offs Local

BSU: National University of Singapore

S*Bio Cordlife KOOPrime etc. Capacity building University–industry linkages Recruit world-class scientists

Figure 2.12

The Biology industry cluster in Singapore

Source: Complied by A. Kuchiki based on Chapter 5.

UILs

IAC BAC

Automobile Cluster in Malaysia 47

2.7 Summary and conclusions In this study we applied the Flowchart Approach to Malaysia’s automobile industry cluster policy, with the aim of elucidating problems in the industrial cluster policy by specifying actors and prioritizing policy measures. We conducted interviews and questionnaires on the problems, and offered prescriptions for the problems by specifying the actors to solve them according to the Flowchart Approach. In step 1, we found the following three facts regarding Malaysia. First, Malaysia has a shortage of domestic demand for cars. Second, it has a shortage of skilled labour. Third, its supporting industries are underdeveloped compared with those of Thailand. In step 2, we defined the following six questions for applying the Flowchart Approach: Is the size of the automobile market of Malaysia sufficiently large? Can Malaysia be a host for export sites? Will foreign automobile firms invest in Malaysia under the present regulations? Will foreign automobile firms invest in Malaysia despite the shortage of unskilled labour? Can Malaysia catch up with Thailand in the supporting industries? Are the controls over foreign capital inflows into suppliers to the national cars sufficiently deregulated? To find the answers of these questions, we interviewed professionals on the problems of Malaysia’s automobile industry. Three results from the questionnaire are as follows: first, firms in Malaysia should establish sites for exports; second, unskilled labour will not be available in Malaysia by further inviting foreign unskilled labour; third, the laws to attract foreign firms in the supporting industries should be deregulated. In step 3, the Flowchart Approach to specify actors and prioritize policy prescriptions led to the following three policy measures: first, Malaysia should reselect anchor firms or international partners with competitiveness and the network of supporting industries; second, organizations in the public, semi-public and private sectors should endeavour to upgrade skilled labour: Third, the central government must commit to and enforce EPA and FTA in ASEAN in order for the Malaysian industry to enhance competitiveness by opening its market. What kind of lessons did the Malaysian automobile clustering policy leave to the Flowchart Approach? Is it possible to foster the automobile industry to any region without the demand condition? Export processing zones and free trade zones in Asia can satisfy the demand condition, since products in the zones can be exported to the world including the US. Fujita (2008), for example, examines the circular causation of relationship between domestic demand creation and clustering. The circular causation also occurs between the creation of supporting industry and clustering (Venables 1996). The Flowchart Approach provides one insight into this problem. The industrial clustering policy of Taiwan, Korea and China since the 1970s was aimed to agglomerate firms in particular regions and to foster strategic industries such

48

Akifumi Kuchiki

as electronics, ICT, automobiles, etc. The success of these policies is a basis of the Flowchart Approach, which attempts to postulate this process. The economic development of these countries was derived by industrial clustering, where anchor firms contributed to produce better and cheaper products which contributed to export as well as to create employment and demand by raising the income level. In addition to these, information flow related to technology, know-how and business models created further incentive to agglomerate firms to clusters. This process forms a positive cycle, not viscious cycle, initiated by the industrial clustering policy. The growth of such clustering contributed to creation of further demand and supporting industries. The Flowchart Approach thus solves the problem of circular causation, which is thought to be a contribution to spatial economics. This chapter attempts to examine and elaborate further the Flowchart Approach from the practical use by taking the Malaysian automobile industry, which suggests local and minimum business conditions of cluster formation. During implementation of the policy, unexpected situations or shocks might occur, but the Flowchart Approach does not cover these matters since it is an ex ante concept and provides a theatrical model to construct a particular cluster policy. They might be critical to the cluster formation. The typical example is coordination failure among actors of the policy such as governments, anchor firms and supporting industries which prevent the development of industrial clusters. Without the feedback process to revise the policy after unexpected matters occur, an industrial clustering policy might just end up with the protection of an inefficient industry. This chapter examines this kind of problem in examining the success or failure of the Malaysian automobile cluster by considering domestic demand and supporting industries, which consist of two major factors to the Flowchart Approach. This analysis provides a lesson not only for elaborating the Flowchart Approach but also for constructing and implementing a practical cluster policy.

References Brimble P. and R. Doner (2007) ‘University-Industry Linkages and Economic Development: The Case of Thailand.’ World Development, Vol. 35, No. 6, pp. 1056–1074. Chen M. and M. Kenney (2007) ‘Universities/Research Institutes and Regional Innovation Systems: The Cases of Beijing and Shenzhen.’ World Development, Vol. 35, No. 6, pp. 1021–1036. Fourin (2006) Monthly Report on the Global Automotive Industry, Nagoyo: Fourin, April. —— (2009) Asia Automotive Industry 2009 Yearbook. Nagoya: Fourin. Fujita, M. (2008) ‘Formation and Growth of Economic Agglomerations and Industrial Clusters: A Theoretical Framework from the Viewpoint of Spatial Economics,’ in The Flowchart Approach to Industrial Cluster Policy, A. Kuchiki and M. Tsuji Eds., Basingstoke: Palgrave Macmillan, pp. 18–37.

Automobile Cluster in Malaysia 49 Hershberg, E., K. Nabeshima, and S. Yusuf (2007) ‘Opening the Ivory Tower to Business: University-Industry Linkages and the Development of KnowledgeIntensive Clusters in Asian Cities.’ World Development, Vol. 35, No. 6, pp. 931–940. JAMA (2005) 2005 Annual Automobile Statistics of the World. Tokyo: JAMA. Japan Bank for International Cooperation (2007) Survey Report on Overseas Business Operations by Japanese Manufacturing Companies. Tokyo: Japan Bank for International Cooperation. Japan External Trade Organization (2006) Current Management of Japanese Manufacturing Industries in Asia. Tokyo: Japan External Trade Organization. Komiya, R., M. Okuno, and K. Suzumura (eds) (1988) Industrial Policy of Japan. London: Academic Press Inc. Kuchiki, A. (2005) ‘The Flowchart Approach to Asia’s Industrial Cluster Policy,’ in Industrial Clusters in Asia, A Kuchiki and M. Tsuji Eds., Basingstoke: Palgrave Macmillan, pp. 169–199. —— (2007) ‘Clusters and Innovation: Beijing’s Hi-technology Industry Cluster and Guangzhou’s Automobile Industry Cluster.’ Chiba: IDE-Japan External Trade Organization, D. P., No. 89. —— (2008) ‘Theory of the Flowchart Approach to Industrial Cluster Policy,’ in The Flowchart Approach to Industrial Cluster Policy, A. Kuchiki and M. Tsuji Eds, Basingstoke: Palgrave Macmillan, pp. 285–311. Kuchiki, A. and M. Tsuji (eds) (2005) Industrial Cluster in Asia Analysis of their Competition and Cooperation. Basingstoke: Palgrave Macmillan. —— (eds) (2008) The Flowchart Approach to Industrial Cluster Policy. Basingstoke: Palgrave Macmillan. Malaysia Trade and Industry (2006) Third Industrial Master Plan 2006–2020. Kuala Lumpur. Markusen, A. (1996) ‘Sticky Places in Slippery Space: A Typology of Industrial Districts.’ Economic Geography, Vol. 72, pp. 293–313. Porter, M. E. (1998) The Competitive Advantage of Nations. New York: The Free Press. Prime Minister’s Department (2006) Ninth Malaysia Plan 2006–2010. Putra Jaya. Tsuji, M., Y. Ueki, S. Miyahara, and K. Somrote (2006) ‘An Empirical Examination of Factors Promoting Industrial Clustering in Greater Bangkok, Thailand.’ 10th International Convention of the East Asian Economic Association, 18–19 November, Beijing. Venables, A. J. (1996) ‘Equilibrium Location of Vertically Linked Industries.’ International Economic Review, Vol. 37, No. 2, pp. 341–359. Wu, W. (2007) ‘Cultivating Research Universities and Industrial Linkages in China: The Case of Shanghai.’ World Development, Vol. 35, No. 6, pp. 1075–1093.

3 Industrial Cluster Development and Innovation in Singapore Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

3.1

Introduction

Among developing economies, Singapore has achieved one of the most impressive economic growth records in the past four decades since her political independence in 1965, averaging 7 per cent GDP growth per annum over the 1960–2006 period. Despite an economic slowdown in 2001–2003 (with a strong recovery in 2004), Singapore’s per capita GDP of US$29,474 in 2006 still stands as the second-highest in Asia, at 67 per cent of the US level (IMD 2006). The rapid economic growth of Singapore has been achieved through continuous industrial restructuring and technological upgrading. In the first decade after independence, growth was led largely by labour-intensive manufacturing. In the two subsequent decades, it was propelled by the growth of increasingly technology-intensive manufacturing activities by foreign MNCs, with high-technology products contributing an increasing share of total value added. The development of Singapore into an increasingly important business, financial, transport and communications services hub in the Asia-Pacific region has provided additional engines of growth since the 1980s. Nevertheless, manufacturing has remained important to the economy, with its share of GDP remaining above 25 per cent for most years in the past two decades. Thus in 2006 27.7 per cent of Singapore’s GDP was contributed by the manufacturing sector, and another 26.5 per cent by ICT and financial/business services. Within the manufacturing sector, the key industries of electronics, chemicals, engineering and the biomedical sciences together accounted for $219 billion (93 per cent) of total manufacturing output. Along with its rapid economic growth, Singapore achieved significant technological capability development. Research and Development (R&D) was minimal until the late 1980s, with a Gross Expenditure of R&D (GERD) to GDP ratio of only 0.86 per cent in 1987, significantly below the norm of advanced countries. Since then, however, R&D investment intensity in 50

Cluster Development and Innovation in Singapore

51

Singapore has increased significantly, with GERD experiencing a 13-fold increase between 1987 and 2006, and the GERD/GDP ratio more than doubling to reach 2.4 per cent in 2006, at parity with the OECD average.

3.2 Conceptual framework for analysing the link between innovation and knowledge-based industrial cluster development 3.2.1

Knowledge-based industrial clusters

A knowledge-based industrial cluster is one that derives significant value creation from advanced knowledge creation and utilization. Both of these aspects are important, requiring both the creation of knowledge-intensive output as well as the use of knowledge-intensive processes in generating this output. Such a cluster will be characterized by knowledge-intensity in every component of the cluster. First, there will be sources of knowledge creation that generate intellectual property (as embodied in patents, copyrights, trademarks, etc.) and tacit know-how (such as skills, tacit knowledge and creativity). The creation of both tangible and intangible know-how takes place in every component of the cluster. For example, the development of technical skills takes place within formal education and training institutions but also within firms and R&D institutions through learning by using and learning by doing. Knowledge creation is not in itself sufficient however, since if the knowledge is not effectively utilized, the system will be left with much under-utilized (or mis-utilized) technological resources, resulting in low returns to the efforts expended in the creation process. Knowledge utilization processes, therefore, are equally important in the cluster, embodied in firms’ operating capabilities and innovation capabilities. Even this on its own is insufficient; there is also a need for knowledge transactions between firms to stimulate the creation and utilization of innovations, such as close interaction between suppliers and buyers or users, or strategic technology alliance between firms. A knowledge-based cluster has a number of different components. First, a knowledge infrastructure is required. This comprises public R&D institutes (PRIs) and universities as the lead generators of knowledge, particularly for fields in basic research, as well as for training manpower that will eventually work in other parts of the cluster. Secondly, linkages to lead users of knowledge are critical. Without such linkages, PRIs and universities run the risk of producing innovations and manpower that are irrelevant to industry. These linkages can include university/PRI-industry R&D collaboration, and high involvement of industry in the design of the training programmes of training institutions. In the early stages of cluster development, lead users may be found overseas rather than in the domestic market, necessitating the formation of linkages with firms and institutions in those

52

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

countries. Thirdly, in order for the cluster to be sustainable, a critical mass of knowledge commercializing/innovating firms is required. Fourthly, the cluster requires supporting industries and services. Such support includes industries which provide industry-specific support to firms in the cluster (e.g., suppliers), as well as companies which provide services required by all knowledge-intensive clusters, such as lawyers and patent agents. Finally, the entire cluster must be supported by a regulatory framework and business environment in which to operate. 3.2.2

Key processes in developing knowledge-based clusters

This chapter adopts a framework which is an application of the flowchart approach to industrial cluster formation (Kuchiki 2005), highlighting several policy measures that are especially important in the context of knowledgebased clusters. In order to create a knowledge-based cluster, each of the following components must be put in place: ●

●

●

Establishment of public knowledge infrastructure, that is, universities and PRIs. This may involve creating new institutions. It may also include restructuring existing institutions, or creating new programmes within them, to give priority to the fields of research and education needed for the cluster under development. Attracting private sector actors to the cluster. This includes both knowledgeintensive/commercializing firms which will form the core of the private sector of the cluster, as well as the supporting services which will surround them. The development of the private sector can take the form of both attracting foreign firms to set up operations in the country though DFI, or by nurturing local firms through incentives and development schemes which will attract firms into the industry and encourage those already in the industry to upgrade their knowledge-intensity. Establishing linkages with lead-user markets. These will commonly involve links to overseas markets, particularly for small or late-entrant economies. Business linkages are needed to expand the market for companies in the cluster, and innovation linkages are needed to give companies access to more advanced products and know-how. Such linkages could take the form of anchoring foreign lead-user firms in the country, and then encouraging intra-firm technology transfer between the parent headquarters and the overseas subsidiaries of a transnational corporation. New entrants will then be able to leverage on the expertise of the early entrants for learning and knowledge transfer, thus facilitating cluster growth. A complementary strategy is to build international links through, for example, international R&D consortia, common technical standards coalition, cross-licensing of technologies, or long-term supplier–buyer relationship.

Cluster Development and Innovation in Singapore ●

●

53

Facilitating knowledge flows and network links among the key actors within the cluster. This will include inter-sector networks, such as between universities/PRIs and private firms (e.g., through technology transfer, joint R&D and training links), as well as creating platforms and mechanisms for inter-firm collaboration within the private sector. Examples of these are R&D alliances and industry consortia. Establishing a regulatory framework/business environment.

The measures described above correspond to the key components in the flowchart approach to industrial cluster policy (Kuchiki 2005). For knowledgebased clusters, innovation, knowledge creation and knowledge transfer are especially important, hence the role of R&D linkages, in particular university– industry linkages, is emphasized. This represents an expansion of the capacity building component of the flowchart approach, with emphasis on infrastructure for innovation and institutions for R&D, including PRIs and universities. 3.2.3 The role of the state in developing knowledge-based clusters The state can play a significant role in facilitating the development of knowledge-based clusters through its policies and investment programmes. This is especially true for economies where the overall business or innovation infrastructure is less well developed – in these cases, the state will play a critical role in cluster development. Moreover, given the diverse strategies that can be adopted in the development of the cluster, the strategic choices eventually chosen by the state can have significant impact on the resulting dynamics of cluster development. Some examples of the strategic choices available to public policymakers include: ●

●

●

Choice of actors to promote: The state can choose to focus on either local or foreign resources in developing the cluster. This includes private firms (attracting foreign firms vs nurturing local firms), manpower (recruiting foreign talent vs developing local talent) and even universities/PRIs (attracting foreign institutions vs establishing local institutions). Timing of entry into emerging technologies: The state chooses when to develop the cluster for emerging clusters and technologies. It can enter the global market while the technology is still new, requiring early- entrant strategies, or it can wait until the market and technology is more mature, necessitating using late-follower strategies. Knowledge infrastructure development : The state chooses between the relative emphasis it gives in developing PRIs vs universities. The dynamics of cluster development will also be influenced by decisions such as the timing of investment in public R&D (e.g., whether the state allows the cluster to first be developed by relying more heavily on private R&D, or whether

54

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

it invests early and aggressively in public R&D), and also the training of R&D scientists and engineers (RSEs) (whether it will focus on training of RSEs directly through public institutions, or allow the private sector to play a greater role). 3.2.4 Knowledge-based industrial cluster development: Upgrading existing cluster vs developing new cluster In developing knowledge-based clusters, the state also faces the choice of upgrading the knowledge-intensity of existing clusters or creating entirely new emerging technology clusters. Regardless, there will be common elements in the strategies used, as the key processes for cluster development are common to both new and existing clusters. However, as the case studies below will illustrate, there will also be distinct differences regarding the specific roles and timing of state involvement depending on the maturity and nature of the cluster to be developed.

3.3 Two case illustrations of knowledge-based industrial cluster developments in Singapore This chapter examines the dynamics of formation of two knowledgebased, innovation-driven industrial clusters in Singapore: the biomedical sciences (BMS) cluster, and the offshore marine engineering cluster. The first represents an emerging technology cluster in the early stage of formation; Singapore had virtually no BMS infrastructure or industry to speak of when the government announced its intention to develop the country into a BMS hub in early 2000. The second represents a more mature cluster that has evolved from an earlier shipbuilding and repair industry-base. From another perspective, the former represents a later-entrant approach by Singapore to ‘catch-up’ with more developed clusters in other advanced countries. The latter cluster in Singapore had already become one of the leading hubs for offshore oil and gas platform production in the world, being the home-base for global leaders like Keppel Offshore and Marine; thus the transformation of Singapore’s maritime services cluster into an International Maritime Centre involved upgrading the knowledgeintensity of existing industries. By comparing and contrasting these two clusters, we highlight the key challenges and relevant policy implications for promoting knowledge-based industrial clusters at different stages of formation. 3.3.1 3.3.1.1

Creating a new cluster: The biomedical sciences (BMS) cluster Development of the BMS cluster in Singapore

For much of its history of rapid economic growth, Singapore had relied on a strategy of attracting DFI from global MNCs and leveraging them to exploit technologies and know-how developed elsewhere (Wong 2001;

Cluster Development and Innovation in Singapore

55

Wong et al. 2005). This global MNC-leveraging strategy has served Singapore well in the past, by making Singapore a leading information technology and electronics manufacturing and services hub in the world (Wong 2002). The same leveraging strategy was adopted in the pharmaceuticals sector, although on a smaller scale and starting later than for IT and electronics manufacturing, and appears to have been similarly effective in turning Singapore into a major pharmaceutical manufacturing hub. As can be seen from Table 3.1a, pharmaceutical manufacturing output in Singapore has grown rapidly since 1980 (18 per cent per annum), reaching S$20.9 billion in 2006. Similarly, its value added has grown at 18.8 per cent per annum over the same period, reaching $12.4 billion in 2006. This amounted to a contribution of 22.4 per cent of total value added in the manufacturing sector, up from 2 per cent in 1980. Reflecting the high capital intensity and scale of operations of such manufacturing activities, the average capital per worker for the industry amounted to S$0.95 million per worker in 2005, while the average output per firm was S$376.9 million, both significantly above the average of all manufacturing. The main drive to create a BMS cluster in Singapore as a whole, however, began in 2000, when the Singapore government announced a strategic shift towards the promotion of biomedical science and technology in order to diversify from high dependence on IT/electronics manufacturing. The intention was for life sciences to become a key pillar of Singapore’s economy, alongside electronics, engineering and chemicals. The government’s vision is to turn Singapore into Asia’s premier hub for biomedical sciences, with world-class capabilities across the entire value chain, from basic research to clinical trials, product/process development, full-scale manufacturing and health-care delivery (Biomed-Singapore 2003). In order to jumpstart the development of the BMS cluster, a coordinated set of major new initiatives was launched (see Table 3.2 for a summary of major initiatives and developments in the Singapore BMS cluster). A US$1 billion fund was initially allocated to boost public investment in several new life science research institutes, to co-fund new R&D projects by global pharmaceutical firms, as well as to initiate the building of a new life science complex called Biopolis. Additional public funding was further announced to sustain the growth of the life science cluster beyond 2006 (Wong 2007). These initiatives, which will be discussed in detail below, have had visible impacts on the biomedical sector in Singapore. The sector has expanded significantly, with pharmaceutical manufacturing output more than quadrupling between 2000 and 2006. Furthermore, a medical technology manufacturing industry has emerged. The output of the medical technology sector has grown from only $31.4 million in 1980 to $2.1 billion in 2006, although it remains only one-tenth the size of pharmaceuticals (Tables 3.1a–c and 3.3). With these two sectors combined, the BMS cluster as a whole had an

Table 3.1a

Year

Profile of the Singapore pharmaceuticals sector, 1980–2006 Number of Number firms employed

1980 1985 1990 1995 2000 2001 2002 2003 2004 2005 2006

17 16 18 18 25 28 38 40 43 43 na

1980–1990 1990–2000 2000–2006 1980–2006

0.6 3.3 11.5 3.8

1,270 1,463 1,645 1,855 1,928 2,375 3,203 3,584 3,857 3,903 4,020 2.6 1.6 13.0 4.5

Net value Output added S$ million S$ million

Fixed asset/ labour $’000

Output/ firm $ million

Val. add/ output %

109.3 227.7 453.0 573.3 1,555.5 1,177.7 1,527.8 1,603.4 2,314.7 2,078.0 3,073

42.9 120.6 117.5 314.7 445.3 878.0 813.9 893.0 960.7 947.4 na

16.8 31.7 56.3 74.6 193.6 183.4 215.0 255.4 362.9 376.9 na

48.7 65.6 73.6 79.2 62.0 54.5 59.9 56.2 57.2 50.0 59.0

Average per annum growth rate (%) 13.5 18.3 13.5 15.3 16.9 14.9 16.1 13.1 27.6 26.6 33.9 12.0 18.0 18.8 18.4 13.7

10.6 14.3 16.3 13.2

12.9 13.1 14.3 13.2

285.1 507.9 1,013.1 1,342.5 4,839.1 5,134.2 8,170.7 10,216.9 15,605.8 16,208.8 20,934

138.8 333.2 745.2 1,063.5 2,999.0 2,797.1 4,893.7 5,746.5 8,927.9 8,110.3 12,355

Net fixed assets S$ million 54.5 176.5 193.4 583.7 858.6 2,085.3 2,607.1 3,200.7 3,705.6 3,697.6 na

Val. add/ labour $’000

Note : Calculated using 2005 data where 2006 data is not available. ‘na’ means data not available. Source: Census of Manufacturing Activities, data from EDB, BMRC (2007), ‘Exceptional Growth for Singapore’s Biomedical Sciences Industry,’ www.biomed-singapore.com/bms/sg/en_uk/index/newsroom/pressrelease/year_2007/6_feb_-_exceptional.html.

Table 3.1b

Profile of the Singapore medical technology sector, 1980–2006

Year

No. of firms

1980 1985 1990a 1995 2000 2001 2002 2003 2004 2005 2006

7 6 14 16 21 24 55 55 58 62 na

1980–1990 1990–2000 2000–2006b 1980–2006b

7.2 4.1 24.2 9.1

No. employed 878 1,192 2,793 3,404 3,952 4,510 4,520 5,058 5,536 6,268 6,551 12.3 3.5 8.8 8.0

Output S$ million 31.4 103.2 293.8 641.8 1,543.0 1,655.4 1,853.2 1,967.1 1,977.0 2,103 2,069 25.1 18.0 5.0 17.5

Net value added S$ million 10.0 75.0 151.9 305.1 820.9 904.9 992.3 1,061.7 884.2 1,115.6 1,210

Net fixed Val. assets add/labour S$ million $’000 na na 189.7 215.9 338.6 303.5 329.5 324.0 342.4 509.7 na

11.4 62.9 54.4 89.6 207.7 200.6 219.5 209.9 159.7 178.0 184.7

Average per annum growth rate (%) 31.3 na 16.9 18.4 6.0 14.3 6.7 8.5 −1.9 20.3 6.8c 11.3

Fixed asset/labour $’000

Output/firm Val. $ million add/output

na na 67.9 63.4 85.7 67.3 72.9 64.0 61.9 81.3 na

4.5 17.2 21.0 40.1 73.5 69.0 33.7 35.8 34.1 36.5 na

na 2.4 −1.0 1.2c

16.7 13.3 −13.1 8.7

31.9 72.7 51.7 47.5 53.2 54.7 53.5 54.0 44.7 49.2 58.5

Notes: na’ means data not available. a Data is for 1991. b Calculated using 2005 data where 2006 data is not available. ‘na’ means data not available. c 1991–2005. Source: Census of Manufacturing Activities, data from EDB, BMRC (2007), ‘Exceptional Growth for Singapore’s Biomedical Sciences Industry,’ www.biomed-singapore.com/bms/sg/en_uk/index/newsroom/pressrelease/year_2007/6_feb_-_exceptional.html.

Table 3.1c

Profile of the Singapore biomedical sciences (BMS) sector, 1980–2006

Year

No. of firms

No. employed

1980 1985 1990a 1995 2000 2001 2002 2003 2004 2005 2006

24 22 32 34 46 52 93 95 101 105 na

2,148 2,655 4,438 5,259 5,880 6,885 7,723 8,642 9,393 10,171 10,571

1980–1990 1990–2000 2000–2006b 1980–2006b

2.9 3.7 17.9 6.1

7.5 2.9 10.3 6.3

Output S$ million 316.5 611.1 1,306.9 1,984.3 6,382.1 6,789.6 10,023.9 12,184 17,582.8 18,311.8 23,003 15.2 17.2 23.8 17.9

Net value added S$ million 148.8 408.2 897.1 1,368.6 3,819.9 3,702 5,886 6,808.2 9,812.1 9,225.9 13,565

Net fixed assets S$ million

Val. add/labour $’000

Fixed asset/labour $’000

na na 383.1 799.6 1,197.2 2,388.8 2,936.6 3,524.7 4,048 4,207.3 na

69.3 153.7 202.1 260.2 649.6 537.7 762.1 787.8 1,044.6 907.1 1,283.2

na na 86.3 152.0 203.6 347.0 380.2 407.8 431.0 413.7 na

Average per annum growth rate (%) 19.7 na 11.3 15.6 12.1 12.4 23.5 28.6 12.0 19.0 17.3c 11.9

na 9.0 15.2 11.0 c

Val. Output/firm add/output $ million % 13.2 27.8 40.8 58.4 138.7 130.6 107.8 128.3 174.1 174.4 na

47.0 66.8 68.6 69.0 59.9 54.5 58.7 55.9 55.8 50.4 59.0

11.9 13.0 4.7 10.9

Notes: ‘na’ means data not available. a Calculated using 1991 data for medical technology. b Calculated using 2005 data where 2006 data is not available. c 1990–2005. Source: Census of Manufacturing Activities, data from EDB, BMRC (2007), ‘Exceptional Growth for Singapore’s Biomedical Sciences Industry,’ www.biomed-singapore.com/bms/sg/en_uk/index/newsroom/pressrelease/year_2007/6_feb_-_exceptional.html.

Cluster Development and Innovation in Singapore Table 3.2 1987 1995 1998

Milestones in the Singapore biomedical sector ● ● ● ●

1999 2000

● ●

● ● ●

● ● ●

2001

●

● ● ● ● ●

2002

● ●

● ●

2003

●

● ● ● ●

2004

● ● ● ● ● ● ● ● ●

●

2005

59

● ●

Set-up of Institute of Molecular and Cell Biology (IMCB) Set-up of Bioprocessing Technology Institute (BTI) Set-up of Centre for Drug Evaluation (CDE) World-renowned Johns Hopkins University set up Johns Hopkins Singapore Set-up of Genetics Modification Advisory Committee (GMAC) Singapore became first Asian country to accede to the Pharmaceutical Inspection Co-operation Scheme, Geneva Set-up of Genomics Institute of Singapore (GIS) Set-up of Life Sciences Ministerial Committee Agency for Science, Technology and Research (A*STAR) established Biomedical Research Council (BMRC) Set-up of Bioethics Advisory Committee (BAC) Set-up of Biomedical Sciences International Advisory Council (IAC) Tuas Biomedical Park Formation of Biomedical Sciences Manpower Advisory Committee (BMAC) Lilly sets up Biology R&D Centre focused on systems biology Set-up of Bioinformatics Institute (BII) Groundbreaking of Biopolis Set-up of Norvatis Institute for Tropical Diseases (NITD) in Singapore Biomedical Sciences Innovate ‘n’ Create Scheme Set-up of Singapore Tissue Network (STN) Merger of Laboratories for Information Technology and Institute for Communications Research to form Institute for Infocomm Research (IIR) Set-up of Institute of Bioengineering and Nanotechnology (IBN) Set-up of Cancer Syndicate Centre for Natural Product Research privatized to become MerLion Pharmaceuticals Launch of SingaporeMedicine Launch of Proof of Concept (POC) Scheme Opening of Biopolis Establishment of Singapore Dengue Consortium Set-up of The Regional Emerging Diseases Intervention (REDI) Centre Set-up of The Centre for Molecular Medicine (CMM) Set-up of Chemical Process Technology Centre (CPTC) Opening of Swiss House Launch of Singapore Researchers Database Passage of the Human Cloning And Other Prohibited Practices Bill Setup of GSK Corporate R&D Centre Launch of BioSingapore The National Advisory Committee for Laboratory Animal Research (NACLAR) announced amendments to the Animals and Birds Act to prevent inhumane treatment of lab animals BAC announced the publication of ‘Research involving Human Subjects: Guidelines for IRBs’ Launch of Medtech Concept Launch of Medtech Local Supplier Group Continued

60

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Table 3.2

Continued ● ●

●

2006

● ●

●

●

●

●

2007

● ● ●

●

●

●

● ●

●

Chemical Synthesis Laboratory at Biopolis opened Government accepts Bioethics Advisory’s paper ‘Genetic Testing and Genetic Research’ Establishment of Singapore Stem Cell Consortium New Bioimaging and Stem Cell Laboratories opened @ Biopolis GMAC releases the Singapore Biosafety Guidelines for Research on GMOs Singapore Government approves S$550 million for Biomedical Sciences Phase 2 New Executive Committee to lead Phase 2 of Biomedical Sciences initiative – Sector to get an injection of S$1.44 billion STaR (Singapore Translation Research) Investigatorship Awards & Translational and Clinical Research Flagship Programme launched Biopolis Phase 2 officially opened Opening of Singapore Institute for Clinical Sciences (SICS) 11 institutions sign MOU for Singapore Dengue Consortium Launch of Singapore chapter of the Association for Clinical Research Professionals (ACRP) Establishment of a Clinical Imaging Research Centre by A*STAR and NUS, with Siemens as industry partner Launch of Facility Sharing Programme under the ‘Growing Enterprises with Technology Upgrade’ (GET-UP) programme, allowing SMEs to have direct access to facilities at 7 A*STAR research institutes Initiatives endorsed by the Biomedical Sciences IAC: ● A*STAR - Duke-NUS GMS Neuroscience Research Partnership ● Development of new research infrastructure at NUS Kent Ridge Campus and Outram Campus ● Proposal to set up a national Academic Clinical Research Organization (ACRO) to provide core services and infrastructure as well as intellectual leadership for later phases of clinical research in Singapore A*STAR - Duke-NUS GMS Neuroscience Research Partnership announced MOU signed Ludwig Institute for Cancer Research to establish a branch for translational and clinical cancer research in Singapore Opening of Institute of Medical Biology (IMB)

Source: A*STAR, Bio-med Singapore.

output of $23.0 billion in 2006, having grown at an average annual rate of 17.9 per cent since 1980. Its fastest growth, however, has been seen since 2000 (23.8 per cent per annum from 2000 to 2006 vs 15.2 per cent from 1980 to 1990 and 17.2 per cent from 1990 to 2000). Similarly, value added in the cluster has grown at an average annual rate of 19.0 per cent between 1980 and 2006 (23.5 per cent from 2000 to 2006) to reach $13.6 billion. Employment in the BMS cluster has almost doubled since 2000, to reach 10,571 in 2006, while labour productivity has also grown steadily, from $0.65 million per worker in 2000 to $1.3 million per worker in 2006.

Cluster Development and Innovation in Singapore

61

Table 3.3 Pharmaceutical and medical technology share of Singapore biomedical sector, 1980–2006 S$ million

Pharmaceuticals Medical technology BMS Total

1980

1990*

285.1 31.4 316.5

2000

Percentage (%) 2006

1980

1990*

2000

2006

Output 1,013.1 4,839.1 20,934 90.1 293.8 1,543 2,069 9.9

77.5 22.5

75.8 24.2

91.0 9.0

1,306.9 6,382.1 23,003

100

100

100

100

83.1 16.9

78.5 21.5

91.1 8.9

100

100

100

100

Employment 4,020 59.1 6,551 40.9

37.1 62.9

32.8 67.2

38.0 62.0

100

100

100

Pharmaceuticals Medical technology BMS Total

138.8 10

745.2 151.9

Value added 2,999.0 12,355 93.3 820.9 1,210 6.7

148.8

897.1

3,819.8

Pharmaceuticals Medical technology BMS Total

1,270 878

1,645 2,793

1,928 3,952

2,148

4,438

5,880

13,565

10,571

100

Note: * Medical technology data is for 1991. Source: Census of Manufacturing Activities, data from EDB, BMRC (2007), ‘Exceptional Growth for Singapore’s Biomedical Sciences Industry,’ www.biomed-singapore.com/bms/sg/en_uk/ index/newsroom/pressrelease/year_2007/6_feb_-_exceptional.html.

The initiatives have also had an impact on R&D. In 2006, BMS R&D in Singapore exceeded S$1 billion, up from $43.1 million in 1993 (Table 3.4). Although this shows R&D growth has been rapid over this time period (averaging 28.3 per cent per annum), it has been particularly so since 2000. From 2000 to 2006 Singapore’s BMS R&D expenditure grew at an average annual rate of 38.2 per cent. The share of total R&D expenditure in biomedical fields has also risen sharply, from less than 5 per cent in the 1990s to over 20 per cent by 2006 (see Table 3.5). However, symptomatic of the long gestation nature of much of biomedical research, the share of biomedical-related patenting in total output of patenting by Singaporebased inventors continued to lag behind its share of R&D spending. As can be seen from Table 3.6, while the cumulative number of biomedical US patents granted to Singapore-based inventors and Singaporebased organizations quadrupled in the eight years 2000–2007 compared to before 2000, the share of biomedical patents in total patents granted remains at 2.6 per cent. Consistent with the larger role of public sector (including universities) in life science research, two-thirds of biomedical R&D expenditure in Singapore in the 2000–2006 period were conducted by public organizations, versus

62 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh Table 3.4 Year

R&D expenditure and manpower in the biomedical sector,a 1993–2006 Private sector

Higher education sector

Government sector

PRIC sector

Total

Total RSEsb

43.1 59.4 81.8 68.5 91.2 118.3 122.9 157.6 310.7 463.1 375.4 760.4 888.9 1,099.5

447 386 570 507 556 625 654 1,333 2,055 2,150 2,504 2,238 2,700 3,049

S$ million 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

3.6 5.0 29.1 7.8 15.0 24.8 37.1 47.0 88.4 147.4 149.3 238.1 312.5 530.7

32.1 39.5 37.0 42.4 47.5 52.1 53.1 62.5 87.3 106.8 87.6 124.9 150.2 178.0

7.4 14.8 15.3 18.2 25.2 35.6 29.1 32.5 57.9 87.5 91.8 116.7 101.4 114.2

0.0 0.0 0.0 0.1 3.5 5.9 3.6 15.6 77.1 121.5 46.7 280.7 324.8 276.6

Notes: a Includes biomedical sciences and biomedical engineering. From 2002 biomedical and related sciences and biomedical engineering. b RSE: No. of full-time equivalent research scientists and engineers. Source: National Survey of R&D in Singapore (various years), A*STAR (previously National Science & Technology Board).

about one third for non-biomedical R&D (Table 3.4). Taking into account the financial incentives given to some private sector pharmaceutical firms to conduct R&D in Singapore, the share of public funding in R&D spending in Singapore is likely to be larger than two-thirds. It is also interesting to note that, while biomedical R&D accounted for only 13 per cent of total research scientists and engineers (RSE) in 2006 (Table 3.5), it accounted for 61 per cent of all PhD RSEs. Again, the larger role of the public sector is not yet reflected in the distribution of patent ownership; ownership of life science patents is fairly equally divided between the public and private sector (about 44 per cent each, with the remainder being assigned to other foreign institutions and individuals, or unassigned) (Table 3.7). Not withstanding this recent rapid growth in importance of life science R&D in Singapore, it is important to recognize that, compared with the advanced countries that are the world leaders in biomedical science and technology, the scale and intensity of Singapore’s biomedical R&D remains modest. For example, Singapore’s total annual biomedical R&D spending of about US$692 million is only a fraction of the US federal annual funding for biomedical R&D (estimated at US$38 billion in 2002). Even in terms of intensity, Singapore’s biomedical share of around 22 per cent of total national R&D is still lower than that of UK and US (over 25 per cent).

Cluster Development and Innovation in Singapore

Biomedicala Shares of Singapore R&D Expenditure and RSEs, 1993–2006

Table 3.5

Biomedical Total R&D Sg R&D expenditure expenditure S$ million S$ million

Year 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

63

43.1 59.4 81.8 68.5 91.2 118.3 122.9 157.6 310.7 463.1 375.4 760.4 888.9 1,099.5

998.2 1,175.0 1,366.6 1,792.1 2,104.6 2,492.3 2,656.3 3,009.5 3,232.7 3,404.7 3,424.5 4,061.9 4,582.2 5,009.7

Biomedical share of total Sg R&D expenditure % 4.3 5.1 6.0 3.8 4.3 4.7 4.6 5.2 9.6 13.6 11.0 18.7 19.4 21.9

Biomedical RSEs

Total RSEs in Sg

Biomedical share of total Sg RSEs %

447 386 570 507 556 625 654 1,333 2,055 2,150 2,504 2,238 2,700 3,049

6,629 7,086 8,340 10,153 11,302 12,655 13,817 14,483 15,366 15,654 17,074 18,935 21,338 22,675

6.7 5.4 6.8 5.0 4.9 4.9 4.7 9.2 13.4 13.7 14.7 11.8 12.7 13.4

Note: a Includes biomedical sciences and biomedical engineering. From 2002 biomedical and related sciences and biomedical engineering. Source: National Survey of R&D in Singapore (various years), A*STAR (previously National Science & Technology Board).

Table 3.6

Share of life science patents in Singapore, 1977–2007 Life science patents

Total Sg patents

Life science patents/ total patents (%)

1977–1999 2000–2007

31 125

1,156 4,795

2.7 2.6

Total

156

5,951

2.6

Year

Note: Singapore assigned patents and patents with at least one Singapore inventor. Following the NBER technological categories, life science patents are taken to be those in drugs, surgery and medical instruments, biotechnology and miscellaneous-drugs and medical. Source: Calculated from USPTO database.

3.3.1.2 Launch of the integrated biomedical sciences (BMS) hub initiative The two arms of the government responsible for establishing the country as a biomedical science hub are the Agency for Science, Technology and Research (A*STAR), formerly known as the National Science and Technology Board, and the Economic Development Board (EDB). A*STAR – or, more specifically, the Biomedical Research Council (BMRC) within A*STAR – concentrates on putting in place the appropriate policies, resources and research and education architecture that will build biomedical science competencies

64

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh Table 3.7

Breakdown of Singapore life science patents by assignee, 1977–2007 No. of patents

Private NUS Government and PRIC Individual/unassigned Other foreign institution Total

68.5 32.5 37.5 13 4.5 156

% 43.9 20.8 24.0 8.3 2.9 100.0

Note: Singapore assigned patents and patents with at least one Singapore inventor. Following the NBER technological categories, life science patents are taken to be those in drugs, surgery and medical instruments, biotechnology and miscellaneous-drugs and medical. Source: Calculated from USPTO database.

internally, including funding and supporting public research initiatives. EDB is responsible for bringing in investments and generating long-term economic value in the BMS sector, which it does primarily through the Biomedical Sciences Group (develops industrial, intellectual and human capital in Singapore in support of the biomedical sciences), and Bio*One Capital (functions as an investment arm). Together, the Biomedical Sciences Group and Bio*One Capital work to attract BMS companies to establish R&D operations in Singapore and develop the local BMS manufacturing sector (Finegold et al. 2004). Figure 3.1 shows a flowchart summarizing the strategies adopted by A*STAR and EDB to develop the BMS cluster in Singapore, while the initiatives themselves are discussed in Section 3.3.1.3. Given the lack of an existing indigenous BMS cluster, Singapore has made extensive use of international talent in its BMS development. The Biomedical Sciences Executive Committee which leads Singapore’s BMS Initiative is advised by the International Advisory Council (IAC), which comprises eminent scientists from around the world, including Sir Richard Sykes (Rector, Imperial College London, UK), Dr John Mendelsohn (President, M. D. Anderson Cancer Center, USA), Dr Alan Bernstein (President, Canadian Institutes of Health Research, Canada), Dr Suzanne Cory (Director, The Walter and Eliza Hall Institute of Medical Research, Australia), Prof. Peter Gruss (President, Max Planck Society, Germany), Dr Philippe Kourilsky (Director, Institute Pasteur, France) and Dr Harriet Wallberg-Henriksson (President, Karolinska Institute, Sweden) Another high-level advisory body is the Bioethics Advisory Committee (BAC), which was formed in 2000, at the time of the US stem cell controversy, to develop recommendations on the legal, ethical and social issue of human-biology research. The recommendations of the Committee, accepted by the government, have led to a regulatory environment in Singapore that is broadly supportive of BMS. The BAC recommended that

Cluster Development and Innovation in Singapore

Develop clinical research in healthcare sector

Promote translational research between healthcare sector & BMS research sector

65

Attract global pharmaceutical MNCs to establish manufacturing operations

Stimulate supporting services

Encourage shift to BMS R&D and clinical trials

Develop Biopolis physical infrastructure to integrate key BMS players

Establish BMS PRIs

Expand BMS R&D and education at local universities

Promote spine-offs & technology commercialization

Attract VCs

Figure 3.1 Singapore’s BMS cluster development strategy

human cloning not be permitted, but doing stem cell research and the use of cloning as a therapeutic tool is allowed.1 This early and clear legal support for stem cell research combines with compliance with strict international guidelines which require seeking consent from couples and using only excess embryos from IVF treatment,2 to give Singapore a relative advantage in stem cell research. Thus the US Institute of Health allows the US federal government to fund research that uses Singapore-produced stem cells (Finegold et al. 2004). 3.3.1.3 BMS hub development: Key elements of development strategy 3.3.1.3.1 Attracting foreign pharmaceutical MNCs into manufacturing, R&D, clinical trials and other knowledge-intensive services. EDB has successfully attracted MNC investments to Singapore; so much so that the BMS cluster is largely dominated by foreign companies. All of the largest pharmaceutical manufacturing firms in operation in Singapore in 2005 are foreign majority owned (see Table 3.8). Firms like GlaxoSmithKline (GSK), Schering-Plough and Merck first came to Singapore to take advantage of the country’s well-established competency in manufacturing. The majority of these firms are headquartered in the United States, and they manufacture pharmaceutical bulk active or intermediate products. Companies such as Genencor, AstraZeneca and BristolMyersSquibb also established

66

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Table 3.8

1 2 3 4 5 6

Top pharmaceutical companies in Singapore, 2005

Company

Nationality

Glaxo Wellcome Manufacturing Pte Ltd Merck Sharp and Dohme Asia Pacific Services Pte Ltd Beecham Pharmaceuticals (PTE) Ltd DSM Nutritional Products Asia Pacific Pte Ltd JMS Singapore Pte Ltd Becton Dickinson Medical (S) Pte Ltd

UK US

Total

UK Netherlands Japan US

2005 sales (S$ million) 30,419.6 2,079.4 562.4 585.1 120.2 114.4 33,881.1

Source: Singapore 1000.

regional headquarters here because Singapore is a major business hub in Asia (Table 3.9). In order to move these MNC investments into higher value-added portions of the biomedical industry value chain, EDB encourages foreign companies to set up R&D or clinical research operations in Singapore. The creation of a sound R&D infrastructure in Singapore, with access to R&D resources from public research institutes and universities has facilitated EDB’s efforts. Some prominent early examples of these partnerships with MNCs include S*Bio, Merlion and Lilly Systems Biology. S*Bio was established as a joint venture between Chiron and EDB using Chiron’s technology platform to develop products for cancer and infectious diseases, especially those in Asia. Merlion originated as a joint venture between Glaxo and EDB to perform more traditional drug-discovery and screening natural samples from across Asia for possible drug targets. After the merger that formed GSK, this unit was spun off and was privatized as a standalone business, with Merlion obtaining all of GSK’s vast library of natural compounds along with its Asian samples. Today, Merlion owns one of the world’s best private collections of natural samples with close to half a million extracts that they are screening for potential drugs, and has grown through collaborations with international drug companies, including Merck, British Biotech and NovImmume. Lilly Systems Biology (LSB) is a wholly owned subsidiary of Lilly that was launched in Singapore in 2002 with generous, multi-year financial incentives from EDB. LSB’s mission is to integrate various biological data and approach the problem of studying complex diseases from a more encompassing perspective of a cell and its system. Through intensive use of computational biology, LSB hopes to discover new drug-targets and biomarkers, and better understand mechanisms of action within the cell (Finegold et al. 2004). An important emerging branch of medical research is ‘translational research’, which is a new approach to the development of drugs and

Cluster Development and Innovation in Singapore

Manufacturing

Table 3.9

67

Major foreign pharmaceutical companies operating in Singapore Date of establishmenta

Size of operation

Company

Business focus

GlaxoSmithKline (GSK)

Bulk active biomanufacturing and regional headquarters

1989

>S$1 billion invested As of 2004, GSK announced an additional $100 million to expand existing manufacturing facility and $50 to develop a Process Technology Centre (to be completed in 2005) 2006: Vaccine manufacturing plant (>S$300 million). To be completed in 2010

Schering-Plough

Bulk active biomanufacturing (Clarityne)

1994 1997: began production

3 of 6 manufacturing plants built in Singapore with a total investment of US$730 million in Singapore, 29,000 employees worldwide

Genset Singapore Oligonucleotide manufacturing

1997

536 employees worldwide (2000) with manufacturing sites in US, Japan and Singapore

Wyeth-Ayerst

Biomanufacturing (Infant nutritional, Premelle)

1999 2002: began operation

Will employ up to 600 employees at full capacity

Aventis

Bulk active biomanufacturing (Enoxaparin)

2000

65,000 employees worldwide

Pfizer

Active ingredient biomanufacturing

2000 (fully operational in 2004)

Merck Sharp & Dohme

Bulk active biomanufacturing (Vioxx, Singulair)

1993b (operational in 2001)

241 employees in Singapore as of mid-2003, 98,000 worldwide 2006: trigeneration facility >200 employees in Singapore, 62,000 employees worldwide 2007: expansion of production facilities, S$100 million Continued

68

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Manufacturing

Table 3.9

Continued

Proligo

Oligonucleotide manufacturing

Novartis

Pharmaceutical production

Ciba Vision

Contact lens manufacturing

Lonza

Biopharmaceuticals production R&D and manufacturing of integrated fluidic circuits Heart valve manufacturing

Fluidigm Corporation

Edwards Lifesciences Corp MDS Sciex

2002 2004: oligonucleotide manufacturing operations 2005: opening of facility (Completion in 2008) 2004: construction of facility began 2006

13 staff

2006

US$250 million

2006

8,000 sq ft facility, will employ up to 500 people (expected) > 30,000 sq. ft facility, will employ >100 people (expected)

S$310 million. Will employ 200 people

Will employ >500 employees

US$250 million

Manufacturing of cellular analysis product, component production for mass spectrometer product lines and the Turbo V ion source Manufacturing of high performance liquid chromatography system Nutritional powder manufacturing

2006 (fully operational by end 2008)

2008

US$280 million

Novo Nordisk

Administration, clinical trial coordination

Quintiles

Clinical research organization

1989: regional HQ 1999: clinical trial centre 1995

50 employees in Singapore, 18,221 employees worldwide 60 employees in Singapore, 800 employees in AsiaPacific, 18,000 employees worldwide Investment of

Waters Corp

Clinical research and trials

Abbott

2006 (1994: Asian HQ)

Continued

Cluster Development and Innovation in Singapore

Research & Development centre

Clinical research and trials

Table 3.9

69

Continued

Covance

Clinical research organization

1996: clinical development 2000: lab site 2006: expanded lab

Pharmacia & Upjohn

Clinical research & medical services

2000

Genelabs Diagnostic

Diagnostic biotechnology

1985

Becton Dickinson

Instrumentation, medical products

Oculex Asia

Ophthalmic drug delivery systems

1986: Medical product 1991: Regional HQ 1995

PerkinElmer

Thermal cycler

1998

Sangui Singapore

Blood supplements and therapeutics

1999

Cell Transplants International Schering-Plough

Cardiac myoblast therapeutics Pilot plant, development laboratories, supplying materials for latestage clinical trials Biomarker discovery tools

2000

Affymetrix

Microarray

2001

Lilly Systems Biology

Systems biology R&D

2001

Surromed

2000

2001

$10 million for clinical trial supplies facility 118 employees in Singapore, 7,900 employees worldwide Investment of >US$1 million See note c

42 employees in Singapore in 2001, 70–90 employees worldwide 1,000 employees in Singapore, 2,500 employees in Asia Pacific 13 employees in Singapore, 59 employees worldwide (mid-2000) Opened S$10 million manufacturing and R&D facility 5–15 employees in Singapore, 30 employees worldwide (mid-2002) 15 employees in Singapore Constructing US$25 million chemical R&D centre

25 employees in Singapore (2002), closed down in 2004 R&D centre is US$25 million investment, 877 employee worldwide (2002) 29 employees in Singapore, 41,000 worldwide Continued

70 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Research & Development centre

Table 3.9

Continued

ViaCell

Stem cell therapeutics

2002

PharmaLogicals Research

Cancers and other diseases in Asians

2002

Novartis

Institute for Tropical Diseases

2002

GlaxoSmithKline

Treatment for neurodegenerative diseases and schizophrenia

2004

Isis Pharmaceuticals

Micro-RNA and antisense drugs to treat Severe Acute Respiratory Syndrome (SARS), cancer and blood diseases New coating and materials technologies Higher order brain functions Chemistry technologies

2004

Essilor

Olympus Corporationd Albany Molecular Research Biosensors International SpineVision

CombinatoRx

Welch Allyn

Development and production of drug eluting stents Develop nextgeneration implants and instruments for non-fusion spinal surgery Development of drug candidates for infectious diseases Development of next generation technology for patient vital signs monitoring

2004

US$4 million invested over a 5-yr period. 138 employees worldwide (2001) S$142 million over the next three years into research on drugs Will house 70 scientists in Singapore, 77,200 employees worldwide Investing S$62 million in preclinical research facility. Will employ 30–35 scientists 12 scientists by end 2005

Current staff strength of 8 people, will grow to 25 staff by 2007

2004 2005

2005

2005

2006

US$20 million, 20 full-time researchers (expected)

Continued

Cluster Development and Innovation in Singapore Table 3.9

Continued

Oxygenix

Development of therapeutic drugs, including small molecules and biologics Biocatalyst R&D for product development

2006

Approx 15 full-time researchers

2007

50 researchers

Ferrosan

Supplements and lifestyle products

Miltenyi Biotec

Magnetic cell separation (HQ Operations) Medical imaging equipment (regional distribution centre)

2003 (2004: conferred International Headquarters Award) 2004

HQ and distribution

Codexis

Philips Medical

Schering AG

Pharmaceuticals

West Clinic Excellence Cancer Center

Cancer treatment and management to patients in Southeast Asia Evaluation and improvement of quality and safety of patient care (Asia Pacific office) Provision of scientific supplies to researchers (S$1 million) Regional clinical management office (approx 10 staff)

Joint Commission International Other

71

Invitrogen Corp

Eisai

2004 (2007: learning centre for medical diagnostics equipment training) 2005 2006

2006

2006

2007

Continued

72 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh Table 3.9

Continued

Other

SGS

PT Kalbe Farma

Lab for quality control testing of pharmaceuticals, biopharmaceuticals and medical devices Research co-ordination and licensing

2006

Notes: a Date established in Singapore. b Establishment date of Sales and Marketing arm. c On 16 April 2003 Pharmacia was acquired by Pfizer. d Olympus partnered Waseda University to establish the Waseda-Olympus Bioscience Research Institute. Source: Updated from Finegold et al. (2004) based on information from Bio-med Singapore website, www.bio-singapore.com and company websites.

treatments, attempting to connect basic research directly to patient care. BMS development in Singapore is moving beyond establishing basic life sciences infrastructure and industry to developing translational and clinical research. Correspondingly, EDB has recently moved to encourage investments in these areas. Examples are The West Clinic’s Excellence Cancer Centre (established in Singapore in 2006) and Eisai’s Regional Clinical Research Centre (2007) (see Table 3.9). 3.3.1.3.2 Developing an integrated physical infrastructure to house the BMS research cluster (‘Biopolis’). Singapore already possesses an excellent general infrastructure (efficient transportation, high speed Internet network, safe and clean city); however it went one step further in building Biopolis, a S$500 million physical hub for life sciences. Dedicated to biomedical R&D activities and designed to foster a collaborative culture among the institutions present and with the nearby National University of Singapore (NUS), the National University Hospital (NUH) and Singapore’s Science Parks, the Biopolis also provides integrated housing and recreation facilities for the many foreign scientists to be attracted to work in the research facilities. Singapore’s seven biomedical public research institutes (PRIs) all have a presence in Biopolis, and are intended to attract biomedical MNCs, start-ups and support services such as lawyers and patent agents to locate there (Finegold et al. 2004). The government hopes that creating such a cluster will create informal networks for knowledge sharing and accelerate the growth of a critical mass of biomedical expertise in Singapore, facilitating its development as a BMS R&D hub for the Asian region. Current private-sector tenants in Biopolis include GSK, Novartis and Isis Pharmaceuticals.

Cluster Development and Innovation in Singapore Table 3.10

73

Establishment of life science public research institutes under A*STAR

PRIC

Established

Institute of Molecular and Cell Biology

1987

Description

Established to help develop and support biomedical R&D capabilities in Singapore. Has core strengths in cell cycling, cell signalling, cell death, cell motility and protein trafficking. Institute of 2002 Founded to conduct research at the cutting-edge Bioengineering of Bioengineering and nanotechnology. Has six and research areas: nanobiotechnology, delivery of Nanotechnology drugs, proteins and genes, tissue engineering, artificial organs and implants, medical devices, and biological and biomedical imaging. Genome Institute 2000 Initially established as the Singapore Genomics of Singapore Program. GIS pursues the integration of technology, genetics and biology towards the goal of individualized medicine. Its focus is to investigate post-sequence genomics; to understand the genetic architecture of panAsian populations with emphasis on cancer biology, pharmacogenomics, stem cell biology and infectious diseases. Bioprocessing 1990 (as Established to develop manpower capabilities Technology Bioprocessing and establish technologies relevant to the Institute Technology bioprocess community. Its core expertise Unit); is in expression engineering, animal cell re-designation technology, stem cells, microbial fermentation, in 2001 product characterization, downstream processing, purification and stability, with supporting proteomics and microarray platform technologies. Bioinformatics 2001 Established to train manpower and build Institute capabilities in bioinformatics. BII’s research focus centres around knowledge discovery from biological data, exploiting high-end computing in biomedicine, advancing molecular imaging of biological processes, modelling of drug design and delivery, computational proteomics and systems biology. Singapore Institute 2007 SICS’s mission is the development of diseasefor Clinical oriented clinical and translational research Sciences programmes in focused disease areas. Institute of 2004 IMB’s mission is to study mechanisms of human Medical Biology (as Centre for disease in order to discover new and effective Molecular therapeutic strategies for improved quality Medicine); of life. It focuses on issues at the interface 2007 between basic science and medicine. The aim is to facilitate the development of translational research by building bridges between clinical and basic science. Source: Websites of individual research institutes/centres.

74 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

3.3.1.3.3 Establishment of seven public R&D institutes (PRI) in BMS. Singapore has seven BMS PRIs, overseen and funded by the BMRC (Table 3.10). Basic research right up to clinical trials are supported using four types of grants: project grants provide seed funding for new investigators; programme grants support more extensive research programmes of established investigators; cooperative grants sponsor inter-disciplinary work; and lastly, core competence grants are meant for PRIs to develop or strengthen capabilities in areas of strategic importance. The oldest of the PRIs is the Institute of Molecular and Cellular Biology (IMCB), first established in 1987 at NUS. Four new PRIs in bioinformatics, genomics, bioprocessing and nanobiotechnology were established over the period 2000–2002, while the existing IMCB was expanded. More recently, the Singapore Institute for Clinical Sciences (SICS) was established to expand Singapore’s clinical R&D capabilities, while the Centre for Molecular Medicine was repositioned as the Institute of Medical Biology (IMB) to facilitate translational research. 3.3.1.3.4 Attracting foreign top talents (‘whales’) and grooming young local talents (‘guppies’). Because of the ambitious scale and speed of development of the BMS cluster in Singapore, the attraction of foreign talent has become an integral part of the government’s life science strategy. Not only would it have taken much too long for the local university to train and develop the large number of scientists needed to staff these major new research institutes, there was also a dearth of local star researchers with sufficient international reputation and stature who can serve as the initial magnet to attract other younger researchers (Zucker and Darby 1996). Consequently, the strategy is to attract a number of internationally renowned scientists (‘whales’) to head research leadership positions in the BMS PRIs, who will in turn recruit more junior scientists from their network contacts as well as attract other young scientists from around the world to work under them. Examples of top scientists attracted to Singapore include: ●

●

●

●

●

Sidney Brenner, a Nobel laureate (Chairman of BMRC and Co-chairman of the IAC); Alan Colman, the leading transgenic animal cloning scientist from Scotland’s Roslin Institute (now Executive Director of Singapore Stem Cell Research Consortium and Principal Investigator at IMB); Edison Liu, the former head of the US National Cancer Institute (now Executive Director of Genome Institute of Singapore); Sir David Lane, the former director of Cancer Research UK’s Cell Transformation Research Group (now Chairman of BMRC, Executive Director of IMCB and CEO, Experimental Therapeutics Centre); Yoshiaki Ito, a leading Japanese cancer researcher (now Principal Investigator of IMCB);

Cluster Development and Innovation in Singapore ●

●

●

●

●

●

●

75

Jackie Ying, a young but rising star researcher from MIT (now Executive Director of the Institute of Bioengineering and Nanotechnology); Axel Ullrich, a well-known molecular biologist from Max Planck Institute for Biochemistry (now Director, Singapore Onco-Genome Laboratory); Dr Birgit Lane, Cox Chair of Anatomy and Cell Biology, University of Dundee (now Executive Director of IMB); Dr Neal Copeland, head of the Molecular Genetics of Oncogenesis Section and Director of the Mouse Cancer Genetics Programme of the National Cancer Institute-Frederick (now Principal Investigator, IMCB); Nancy Jemkins, head of the Molecular Genetics of Development Section of the National Cancer Institute-Frederick (now Principal Investigator, IMCB); Dr Judith Swain, Dean for Translational Medicine and the Founding Director of the College of Integrated Life Sciences at University of California at San Diego (now Executive Director, Singapore Institute of Clinical Sciences); Markus Wenk, a noted biophysicist and lipid researcher from Yale (now Associate Professor at NUS).

In addition to these star scientists, the government is sending the top students from Singapore’s education system (‘guppies’) to leading research universities for graduate science and business education. The government pays for their education provided that they return to Singapore when they complete their studies. The scholarships, which are provided by A*STAR, target different segments of young talent, ranging from those seeking undergraduate and postgraduate studies to medical doctors seeking training to become clinician scientists. As of 2007, more than 100 students have been trained, with a target of 1,000 trained PhDs by 2015. In the long term, the government hopes that Singapore’s own universities and research institutes, bolstered by alliances they have established with universities such as Johns Hopkins and MIT, can grow their own bioscience manpower (Finegold et al. 2004). 3.3.1.3.5 Nurturing venture capital and promoting dedicated biotechnology firms (DBFs) and biomedical device start-ups. Collectively, the measures to create a vibrant BMS R&D landscape in Singapore are intended to lead to promotion of technology commercialization activities through start-ups and spin-offs from the PRIs and universities, as illustrated in Figure 3.1. The availability of venture funding for technology-based firms is an important factor to support such activities. As with other aspects of the BMS cluster in Singapore, the government has played a significant role in developing a specialized BMS venture capital industry. The biomedical science industry is very capital intensive and

76 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

risky, with most new ventures failing to generate a return on their investment. With no history of home-grown, high-tech companies, Singapore has not developed a community of venture capitalists or other private investors who are knowledgeable about and interested in investing in biomedical start-ups. Those investors interested in this sector have tended to put their resources in US firms with more proven track records and lower perceived risk. Consequently, the government took a lead role in directly running a number of life-science related funds, which were subsequently centralized under one fund management umbrella called Bio*One Capital. Bio*One Capital manages about S$1.2 billion in funds, and invests in drug discovery/development, biologics and cellular therapy, and medical technology companies and start-ups. It currently has over 60 companies in its investment portfolio, many of them originally founded outside Singapore. Through a strategy of investing in companies that will bring key new technologies and generate higher value-added research jobs in Singapore, the fund had been instrumental in causing some companies to move some of their operations to Singapore. One example is S*Bio. EDB offered a multimillion- dollar deal to transfer Chiron’s technology platform to a new drug discovery start-up in Singapore, in which Chiron was then given a significant ownership stake. Another example is Bay Areaheadquartered Fluidigm, which chose Singapore to locate her first Asian manufacturing operation (Wong 2007). In addition to bringing overseas BMS investments to Singapore, the efforts of Bio*One and other support mechanisms have resulted in a fledgling dedicated biotech firms (DBFs) sector emerging in Singapore (Table 3.11). This includes a number of spin-offs from local universities (see Table 3.12 for DBF spin-offs from NUS). The main focus areas for Singapore’s DBFs are drug discovery and development (e.g., S*Bio, established in 2000, Merlion Pharmaceuticals (2002), ProTherapuetics (2004)), medical devices (e.g., BioSensors (1990, listed in 2005), Merlin Medical (2002)), stem cells (ES Cell International (2000); Cordlife (2002)) and bioinformatics (KOOPrime (2000); HeliXense (2000); ReceptorScience (2000)). Although the number of DBFs in Singapore is still relatively small when compared to the leading biotech clusters in the world, the record is actually creditable, given that there were virtually no such DBFs before 2000. Indeed, until recently, only two drug-discovery companies, Lynk Biotechnology and AP Genomics, had products. Some still only have drug candidates in their pipeline or remain in the extremely early stages of research. All the bioinformatics firms have some products in the market because product development in software is significantly shorter than it is for drugs. Nevertheless, a major commercial success has yet to emerge among Singapore’s DBFs.

Cluster Development and Innovation in Singapore

Medical devices

Drug discovery

Table 3.11

77

Dedicated biotechnology firms (DBFs) founded in Singapore Products/services

Date of establishment

Company

Business focus

S*Bio

Drug discovery

na

2000

Lynk Biotechnology

Drug discovery and development

Biolyn™ Hair Serum Several leads

2000

APGenomics

Genomics-based products, services, and technology

Dengue SmartPCR™ diagnostic kit

2000

Qugen

Gene therapy

3 products nearing clinical trials

2001

Agenica

Breast cancer treatment

na

2001

Merlion Pharmaceuticals

Drug discovery

na

2002

ProTherapeutics

Sublingually delivered peptide therapeutics

Analgesic peptide

2004

SingVax

Prophylactic vaccines for infectious diseases prevalent in the Asia-Pacific region

Japanese Encephalitis (JE) vaccine for the prevention of JE virus infection and an Enterovirus 71 (EV71) vaccine for the prevention of Hand, Foot and Mouth Disease (in development)

2005

Biosensors International

Minimally invasive surgical devices

Catheters, transducers, stents

1990

SiMEMS Pte Ltd

Manufacture and marketing of MEMS based sensor products that support clinical diagnostics.

Chip-based system that accelerates and miniaturizes the process of extracting and detecting DNA

1999

Forefront Medical Technologya

Medical devices used in anaesthesia

Contract manufacturing

2000

Attogenix Biosystems

Integrated molecular High throughput analysis chip reaction array with microfluidic biochip (AttoChip) technology and real time

2002

Continued

78

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Table 3.11

Continued

Company

Business focus

Products/services

Date of establishment

Stem Cell

Medical devices

sequence detection system (AttoCycler) (to be launched in 2005) Merlin Medical

Minimally invasive medical devices

Coronary stent

2002

Veredus Laboratories

Diagnostic assays for infectious diseases prevalent to the Asia region

VereFlu (TM) Lab-on-Chip to diagnose all flu types (including avian flu)

2003

Amaranth Medical

Bio-absorbable stents for peripheral and coronary vascular applications

Bio-absorbable urinary stent and vascular stents

2005

HealthSTATS International

Continuous, noninvasive biomonitoring devices for hypertension control

BPro®, a medical device allowing hypertension to be analysed at both the macroscopic and microscopic levels

2005

Curiox Biosystems

Commercialize and develop the DropArray technology

DropArray – miniaturized aqueous bioassays for drug discovery and other life sciences applications.

2007

Promatrix Biosciences

Hematopoietic stem cell (HSC) therapy

Has developed a method of constructing multi-cellular tissue to resemble normal tissue

2002/3

ES Cell International

Stem cell research and production

Human embryonic cell lines

2000

CordLife

Stem cell banking & therapeutics

Sample collection and counselling services

2001

Continued

Cluster Development and Innovation in Singapore

Others

Bioinformatics

Table 3.11

79

Continued Date of establishment

Company

Business focus

Products/services

ReceptorScience

Bioinformatics software

ReceptoMiner ™ database and tools

2000

KOOPrime

IT solutions for life sciences

Platforms, engines, databases

2000

HeliXense

Bioinformatics platform

Genomics Research Network Architecture

2000

Mycosphere

Fungal diversity, bioprospecting

Contract research

1997

AP Metrix

Healthcare and patient monitoring

RemoteC@re™ (patient-provider management platform)

2000

BioSurfactants

Surfactants manufacturing & development

na

2001

Maccine

Preclinical service provider supporting drug development and safety assessment

GLP laboratory facility will be ready by Q1 2006

2003

A-Bio Pharma

Biologics contract manufacturer

Partnership with GlaxoSmithKline to develop and manufacture vaccines

2003

NeuroVision

Non-surgical treatment for lowgrade myopia and amblyopia

Neural Vision Correction technology

Davos Life Science

Active ingredients based on natural tocotrienols

Tocotrienol isomers, health supplements Co3E and Co3E + CoQ10

2004 treatment centre 2004

Note: a Forefront Medical Technology is a 50–50 joint venture between Singapore company Vicplas International and UK-based Laryngeal Mask Company. Source: Updated from Finegold et al. (2004), ‘Adapting a Foreign-Direct Investment Strategy to the Knowledge Economy: The Case of Singapore’s Emerging Biotechnology Cluster,’ based on information from Bio-med Singapore website www.bio-singapore.com and company websites.

Table 3.12

Profile of NUS biomedical-related spin-off companies

Incorporated date

Name

Nature of business

Founders from NUS

Department

1995

Allegro Science Pte Ltd

Dr Victor Wong Wong Thi

Biology

1999

BioMedical Research and Support Services BioNutra International Pte Ltd

Produce and market sequencing and DNA labelling kits Biological evaluation of medical devices and equipment Nutraceuticals and biopharmaceuticals

Assoc. Prof. Eugene Khor

Chemistry

Mr Victor Ong Yek Cheng, Assoc. Prof. Paul Heng, Assoc. Prof. Yong Eu Leong Prof. Ching Chi Ban, Assoc. Prof. Ng Siu Choon

Pharmacy

Scientific Adviser, Dr Ariff Bongso Mr Lim Teck Sin

Medicine – Obstetrics and Gynaecology Center for Natural Product Research Medicine – Physiology Medicine Bio-engineering

2001

2002

Chiral Sciences & Technologies Pte Ltd

2000

ES Cell Pte Ltd

2000

KOO Prime Pte Ltd

2000 1997 2003

LYNK Biotech Pte Ltd Oribiotech Pte Ltd OsteoPore Pte Ltd

2003

Quantagen Pte Ltd

2003

BioNano International Singapore Pte Ltd

2004

ProTherapeutics

Sublingual delivery of peptide therapeutics

2005

BioMers

Polymer composite products for numerous biomedical applications (orthodontics, dentistry, orthopedics)

Source: National University of Singapore.

Pharmaceutical and biopharmaceutical intermediates and fine chemicals Embryonic stem cell technology IT solutions provider for the life sciences, bio-mining software Drug development technology Develop tumour markers Biodegradable Bone Scaffold Label-free detection technology for gene analysis Biosensors using multi-walled carbon nanotubes

Assoc. Prof. Lee Chee Wee Dr Ng Wee Chit Prof. Teo Swee Hin Dr Dietmar Hutmacher Prof. Casey Chan Assoc. Prof. Sheu Fwu-Shan, Dr Ye Jian-Shan, Assoc. Prof. Lim Tit Meng Prof. Manjunatha Kini, Assoc. Prof. Ge Ruowen, Assoc. Prof. Peter Wong Ms Renuga Gopal, Ms Karen Teo, Dr Mervyn Fathianathan

Chemistry

Orthopaedic Biological Sciences

Biological Sciences, Pharmacology Mechanical Engineering, Restorative Dentistry

Cluster Development and Innovation in Singapore

81

3.3.1.3.6 Expanding clinical research capabilities in the health-care services sector. The first phase of Singapore’s BMS development (2000–2005) focused on establishing a foundation of basic biomedical research in Singapore. From 2006, Singapore’s BMS Initiative moved into its second phase, which focuses on the development of capabilities in clinical and translational research, while continuing to strengthen basic sciences (A*STAR 2007b). Singapore has several advantages in clinical research, including a good health-care system. There are seven public hospitals, six national speciality centres on cancer, cardiac, eye, skin, neuroscience and dental care, and 16 private hospitals. It also has a primary health-care network of 18 public polyclinics and over 2,000 private medical practitioners. The country’s health-care system has a strong regional reputation for advanced services, attracting over 370,000 foreign patients in 2005 and having a target of 1 million by 2012. Other advantages include its compact size and a population with a mix of Asian ethnic groups, which makes it conducive for developing new treatments and technologies, as well as drug trials customized to Asian populations. Earlier in its BMS cluster development, Singapore began to market itself as a regional clinical trials centre. It had some early success attracting companies like Pharmacia, Novo Nordisk and some of the large contract research organizations (CROs) to establish clinical trial centres in Singapore. Johns Hopkins University and NUH set up an International Medical Centre to provide patient care and to conduct clinical trials in oncology. This centre also offers clinical education programmes and degrees in conjunction with NUS. Despite these early successes, there are still relatively few clinical trials taking place in Singapore. This may be due to the strong and growing competition for the Asian clinical research market from Taiwan, Australia and Japan, which have the advantage of larger domestic markets and the fact that pharmaceutical companies may be reluctant to use these clinical trial centres due to an unproven track record (Finegold et al. 2004). More recently, further development of clinical research capabilities has gained prominence. To this end, clinical research was explicitly included in the mandate of the Ministry of Health (MOH) as of 2006, and the National Medical Research Council (NMRC) under MOH has significantly increased funding for clinical research by the health-care sector. The establishment of the SICS was a further step in this direction, with the institute focusing on the development of disease-oriented clinical and translational research programmes in the areas of genetic medicine, hepatic diseases and metabolic diseases. The development of clinical research necessarily involves developing translational research; thus the SICS and the other newest PRI, IMB, both have explicit aims of building bridges between basic science conducted by the other PRIs and clinical research programmes in Singapore’s public hospitals, disease centres and the universities.

82 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

As with other sectors within the BMS cluster, clinical research suffers from a lack of manpower; as such, the development of clinician-scientist manpower has also come into focus. One scheme is the National Science Scholarship offered by A*STAR which specially targets medical doctors who want to become clinician scientists. Two other talent attraction and development schemes recently launched by A*STAR are the Singapore Translational Research Investigatorship, which aims to recruit world-class clinician scientists and clinician investigators to undertake clinical and translational research in Singapore, and the Clinician Scientist Award, which targets top local clinicians with proven leadership potential towards research (A*STAR 2007b). 3.3.1.3.7 Promoting translational research and other linkages between R&D institutes, universities and the health-care services sector. In the development of clinical and translational research capabilities, there has been a deliberate strategy of forming and promoting collaborations between clinicians and scientists in multiple agencies. Recent examples of this are the consortia initiated by BMRC to promote translational research links between the BMS PRIs/universities and the health-care sector. These consortia are to coordinate and drive translational research at the national level by consolidating existing research activities and scientific expertise, optimizing the use of critical research resources, filling in gaps in research capabilities and building sufficient critical mass to make Singapore’s efforts in translational research internationally competitive. The consortia that have been set up to date include: ● ● ● ● ●

Singapore Cancer Syndicate; Singapore Bio-imaging Consortium; Singapore Stem-Cell Consortium; Singapore Consortium of Cohort Studies; Singapore Immunology Network.

These consortia engage in a variety of activities including funding of joint projects, engaging in joint training and establishment of research infrastructure and links between local and overseas institutes. Singapore has also become a member of the International Biomarker Consortium headed by Dr Leland Hartwell, President and Director of the Fred Hutchinson Cancer Research Center in Seattle. It aims to launch a large-scale, coordinated effort to discover biomarkers for the early detection of cancer by providing opportunities for collaboration and platforms for data and technology and sharing among consortium members. Singapore’s project focuses on biomarkers for gastric cancer, and involves a collaborative research effort NUH, NUS, the Genome Institute of Singapore (GIS), IMCB and the Bioinformatics Institute (BII). There are also plans to establish a Biomarkers Consortium within Singapore, which would bring together

Cluster Development and Innovation in Singapore

83

basic and clinical research groups working on biomarker discovery in cancer and other diseases (A*STAR website). Although it is too early to evaluate the effectiveness of the consortia, their formation is to be welcomed. One weakness of the Singapore innovation system as a whole is a lack of linkages between sectors; the expansion and promotion of collaboration through consortia is thus a step in the right direction. 3.3.1.4 Growing R&D and the educational role of local universities 3.3.1.4.1 Education The development of the BMS cluster in Singapore has necessitated a growing role played by local universities in both R&D and education. In terms of education, the universities have put in place programmes to help develop the manpower necessary for the BMS cluster. Among the initiatives undertaken by NUS, the most ambitious has been the building of a second medical school. Unlike the existing medical school, which is in the British tradition of taking students directly from high schools, the new school was modelled after the US postgraduate, professional medical school, with students drawn from various disciplines and faculty recruited to emphasize research excellence. The school was established in collaboration with a leading US medical school (Duke University), and is located on the same campus as the largest public hospital (Singapore General Hospital (SGH)) to facilitate close interactions, particularly in research (Wong 2007). NUS also set up an Office of Life Sciences (OLS) in 2001 to integrate and facilitate life sciences research and education throughout the university and its affiliated institutions. Among the educational initiatives introduced by OLS is an integrated life science undergraduate major programme that involves the participation of five core faculties (computing, dentistry, engineering, medicine and science). OLS also established a bioinformatics programme with faculty members from the computing and medical schools, as well the engineering and science faculties. A new bioengineering division was also set up in the engineering faculty in 2001. The division has been structured so as to cross traditional departmental boundaries within the faculty and the university. Thus policy decisions are made by a board comprising academic staff from the engineering, science and medicine faculties. Similarly, academic staff in the division hold joint appointments with other departments in the engineering faculty and other faculties and research institutes. This gives students to access expertise and resources across the faculties involved; for example, exposure to clinical practices in the medical school and NUH (Wong 2007). Singapore’s second-largest university, Nanyang Technological University (NTU), has also entered BMS education and research. Its School of Biological Sciences was established in 2001, within the College of Sciences, while a School of Chemical and Biomedical Engineering has been established within the College of Engineering.

84 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

3.3.1.4.2 Research As mentioned above, BMS research programmes in NUS are consolidated by the Office of Life Sciences. OLS brought together researchers from the five core faculties involved in life sciences to identify and agree on ten areas of research, grouped under two broad headings of diseases (cancer, neurobiology/ageing, vascular biology/angiogenesis, hepatology, infectious diseases) and platform technologies (bioinformatics/registries/molecular epidemiology, structural biology/proteomics/genomics, immunology, bioengineering, experimental therapeutics/medicinal chemistry/toxicology/clinical trials). The forging of this consensus provides greater focus for BMS research within NUS, and for new collaborations established with other research institutes in Singapore and overseas (Wong 2007). The National University Medical Institutes (NUMI), formed in 1994, translates some of these target research areas into specific research programmes. It has research programmes in oncology, low temperature preservation, neurobiology, diabetes and ROS biology and apoptosis. NUMI also supports research in the medical school, providing centralized research facilities and services such as confocal microscopy, DNA sequencing and flow cytometry. The Proteins and Proteomics Centre similarly supports BMS research within the NUS Faculty of Science. The Centre provides equipment and services for researchers, as well as training for both undergraduate and postgraduate students in support of research projects in proteomics, protein science and technology, chemical biology, drug design, immunology and structural biology. Within the engineering faculty, the Division of Bioengineering has set up a bioengineering and nanobioengineering corridor to facilitate research in nanotechnology and bioengineering. The corridor consists of numerous multidisciplinary laboratories in a single location, with the aim of encouraging cross-fertilization of ideas and multidisciplinary teaching and research. The laboratories include those on biomaterials, biomechanics, biophysics, bionanotechnology, biosignal processing and instrumentation, chemotherapeutic engineering, computational bioengineering, nanobioengineering, nano biomechanics, tissue engineering and biofluids. In addition to research within the various faculties of universities, BMS research is conducted in university-affiliated research institutes. One example is the Temasek Life Sciences Laboratory (TLL), which is affiliated with both NUS and NTU. Located within the NUS campus, TLL and conducts research in molecular biology and genetics. 3.3.1.4.3 Translational research The NUS-Duke Graduate Medical School (GMS) was established in 2005, having a research-intensive curriculum based on the Duke University model of medical education. This focuses on developing clinician scientists who will engage in translational research. As mentioned above, this

Cluster Development and Innovation in Singapore

85

is facilitated by the school’s location on the same campus as the Singapore General Hospital and the national specialty centres. Researchers have access to specialized research support facilities, and the physical proximity facilitates collaboration for translational research. Recently, GMS announced a partnership with A*STAR to develop an integrated multidisciplinary neuroscience research programme with a strong focus on translational research. The Neuroscience Research Partnership will widen the collaborative scope of researchers in GMS further still, to include the resources and expertise within the BMS public research institutes under A*STAR. Another high-profile avenue of translational research is the Singapore Gastric Cancer Consortium, which is a collaboration of researchers from NUS, NUH and the National Cancer Centre Singapore (NCCS), as well as from the PRIs and overseas cancer centres. This research is the first project to be supported by the NMRC’s Translational Clinical Research Flagship Programme, receiving S$25 million over the next five years. The NUS module is built around ongoing clinical studies and relevant research laboratories in the university. The NCCS module will focus on the genetic analysis of gastric cancer, while the NUH module will focus on clinical trials. The consortium has partnered with several international groups as well as with other hospitals, PRIs and universities within Singapore (A*STAR 2007a). The infrastructure for translational research in the university medical schools, and the hospitals with which they are co-located, is being further developed through a $140 million project announced in 2007. The plans include the development of new research buildings for laboratory research, investigational medicine units, and animal research facilities. Not only will research facilities be expanded, but the research infrastructure of the institutions involved will be integrated, resulting in two campuses for translational research: one comprising NUS and NUH; the other comprising GMS, SGH and the national disease specialty centres (A*STAR 2007b). 3.3.1.4.4 International multidisciplinary research A initiative for promoting international multidisciplinary research among Singapore universities was put in place by the National Research Foundation’s Campus for Research Excellence and Technological Enterprise (CREATE). CREATE aims to foster joint multidisciplinary research and linkages between top research universities and Singapore-based knowledge organizations. The first research centre within CREATE is one that builds on an existing relationship between MIT, NUS and NTU known as the Singapore-MIT Alliance (SMA), in which the three universities collaborate in graduate education and research. The universities have extended their partnership to form the Singapore-MIT Alliance for Research and Technology (SMART) Centre, a research centre where faculty, researchers and graduate students from MIT collaborate with those from universities, polytechnics, research institutes and industry in Singapore and Asia. SMART will establish five

86

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

interdisciplinary research groups, two of which are currently operational. The first is the Centre is on Infectious Diseases, with eight senior MIT faculty members and 17 Singapore collaborators from NUS, NTU, TLL and other PRIs in the fields of biology, engineering, medicine and computing. The second research group is a Centre for Environmental Sensing and Modelling, involving collaborators from MIT, NUS, NTU, the Tropical Marine Science Institute and the Public Utilities Board in the fields of engineering, earth and atmospheric sciences, architecture and computing (NRF 2007). 3.3.2

Offshore marine engineering cluster

The development of Singapore’s offshore marine cluster needs to be examined in the context of the holistic strategies to develop Singapore’s maritime cluster (SMC). Singapore’s maritime cluster has been well established for many years as the island’s strategic location has propelled it to its current position as one of the world’s most important port and shipping locations. The current policy impetus is to enhance the SMC in order to position Singapore as a leading international maritime centre in the Asian region. 3.3.2.1 Overview of the development of Singapore’s maritime cluster (SMC) Singapore’s maritime cluster is defined to broadly comprise two groups: core maritime sectors which include the traditional water transportation sectors and non-core maritime sectors which include services that support marine transportation. Core maritime sectors are those that derive their revenues entirely from maritime-related activities. Non-core sectors are those for which maritime activities form only a part of their total operations. Table 3.13 shows the sectors within these two broad groupings. Table 3.13

Key component industries within Singapore maritime cluster

Traditional ‘core’ maritime sectors

Non-core maritime sectors

Offshore Shipbuilding and repair Shipping lines Ship brokering and chartering Port sector Wholesale and retail marine equipment Classification societies and marine surveying services Shipping agencies Ship management services Ship chandlers Ship bunkering Cruise Inland water transport

Logistics and supporting services, incl.: Freight forwarding Cargo survey services Container services

Source: Wong et al. (2004).

Engineering and other technical services

Ancillary services, incl.: Maritime legal services Maritime finance Maritime insurance Maritime education & training Maritime R&D & technology

Cluster Development and Innovation in Singapore

87

The SMC has experienced impressive growth over the past few years, as seen in Table 3.14. In 2005, the SMC generated value added of $14.3 billion, accounting for 7.4 per cent of Singapore’s total GDP. This compares to 5.1 per cent share of GDP in 2000. In the intervening five years, the SMC grew in excess of 12 per cent annually, while generating employment growth of close to 7 per cent annually. Labour productivity has correspondingly improved over the years. Value added per worker increased from S$117 in 2000 to S$149 in 2005. The SMC’s labour productivity figures compare favourably against those achieved economy-wide, which stood at S$74 in 2000 and S$84 in 2005. Table 3.15 contrasts the economic contribution of the SMC with the maritime clusters of other selected countries. This illustrates the significance of maritime-related activities to Singapore’s economy. Among the different countries included in this benchmarking exercise, Singapore was the only one Table 3.14

Growth trends in Singapore maritime cluster

Direct VA (S$ million) Direct VA as share of GDP (%) No. of employed VA/worker (S$’000) VA/worker – Economy-wide (S$’000)

2000

2005

8,104 5.1 69,257 117 74

14,311 7.4 96,136 149 84

CAGR % (2000–2005) 12.10 na 6.90 4.10 2.10

Note: ‘na’ means data not available. Source: Wong et al. (2007).

Table 3.15

Maritime clusters value added, international benchmarks 2001 Value added in USD (2001)

Denmark Germany Netherlands Norway UK Hong Kong Singapore

Maritime VA excluding non-transportation services (e.g., legal, finance) 1.9 9.7 4 4.8 8.6 3.96 4.3

VA as share of GDP (%)

Total maritime VA including services

Maritime VA excluding non-transportation services (e.g., legal, finance)

Total maritime VA including services

na na na na 10.3 4.04 4.4

1.1 0.5 1 2.9 0.6 2.41 5.2

na na na na 0.7 2.44 5.3

Notes: For all European nations except the UK, maritime VA excludes non-transportation services such as finance, legal services, insurance. Estimates for European countries apart from the UK are average over 1999–2001. For the UK, Hong Kong and Singapore, VA is for 2001. Singapore’s maritime VA includes offshore sector. ‘na’ means data not available. Source: Wong et al. (2004).

88 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh Table 3.16

Linkages of maritime cluster to economy Output

Value added

Estimated multiplier, 2005 (based on 2001 I-O tables) Direct contribution of maritime cluster Indirect contribution of maritime cluster

1.39

0.23

S$ 79.71 billion S$ 31.41 billion

S$14.31 billion S$ 3.45 billion

Total

S$ 111.11 billion

S$ 17.76 billion

Source: Computed from sectoral multipliers provided by Department of Statistics.

where maritime activities account for over 5 per cent of Gross Domestic Product (GDP). The contribution in the other countries was below 3 per cent, even in traditional maritime nations such as Hong Kong and Norway. Additionally, the SMC demonstrates strong linkages to the rest of Singapore’s economy (Table 3.16). In 2005, the SMC generated S$31.4 billion of output in non-maritime related sectors. The SMC also generated $3.45 billion of indirect value added in the rest of the economy, bringing the total (direct + indirect) value-added contribution of the SMC to SGD 17.76 billion or around 9.2 per cent of GDP. Table 3.17 shows the composition of value added and employment in the SMC for 2005. Two sectors, shipping lines and brokering and chartering, account for almost half the value added generated by the SMC, with the ports sector contributing another 14.5 per cent. However, of these three dominant sectors, only brokering and chartering has recorded significant growth over the past five years. Outside of these traditional shipping and port activities, the offshore and marine engineering sectors are significant components of Singapore’s maritime cluster. These sectors account for almost one-fifth of value added in the SMC and generate one-quarter of maritime employment. The offshore construction sector contributed 6.8 per cent of maritime value added and 9.2 per cent of maritime employment. The shipbuilding/repair sector contributes one-tenth of total maritime VA and is the single largest employer of maritime labour, accounting for 38.2 per cent of employment in the SMC. Additionally, the offshore and marine engineering sectors boast of high growth rates in the past five years. For both value added and employment generation, offshore and marine engineering sectors recorded double-digit growth rates per annum between 2000 and 2005. Growth rates in value added generated are especially impressive, outstripping many other sectors in the SMC, which recorded average growth of 6.8 per cent per annum in the same period. 3.3.2.2 Competitiveness and development of the Singapore maritime cluster 3.3.2.2.1 Upgrading of knowledge intensity of existing industries in the maritime cluster The continued growth of Singapore’s maritime cluster has been achieved in the face of increasing regional and global competition. This may be

Table 3.17

Principal statistics of Singapore maritime sector, 2005 Value added 2005 ($ million)

Core maritime sectors Offshore Shipbuilding and repair Wholesale/retail of marine equipment/ accessories Ship chandlers Ship bunkering Shipping lines Cruise Inland water transport Ship brokering services and ship chartering Classification societies and marine surveying services Shipping agencies Port Ship management services Non-core maritime sectors Logistics and supporting services Engineering & other activities Legal services (maritime) Insurance, reinsurance and P&I (maritime) Maritime-related finance Training/education (maritime) Maritime-related R&D and information technology Singapore maritime sector

2005 % share

Employment 2000–2005 CAGR (%)

2005 No.

2005 % share

2000–2005 CAGR (%)

77,036 8,838 36,688 4,063

80.13 9.19 38.16 4.23

12.017 13.3 11.9 6.5

12,865.00 978.0 1,479.9 318.1

89.89 6.83 10.34 2.22

6.335 12.8 10.1 4.5

117.6 266.0 3,479.1 2.0 284.7 3,219.8 152.7

0.82 1.86 24.31 0.01 1.99 22.50 1.07

8.6 14.0 −1.0 14.0 −2.3 21.7 6.8

1,950 1,238 2,573 79 1,916 2,359 901

2.03 1.29 2.68 0.08 1.99 2.45 0.94

12.5 42.3 12.7 −1.0 2.1 32.0 14.1

312.0 2,074.1 181.0 1,446.9 1,017.9 120.1 28.4 86.3 63.2 37.6 93.3

2.18 14.49 1.26 10.11 7.11 0.84 0.20 0.60 0.44 0.26 0.65

−6.0 0.8 −1.7 8.7 6.6 4.5 3.5 2.3 14.5 16.0 2.8

3,542 10,400 2,489 19,100 15,642 859 240 398 282 570 1,109

3.68 10.82 2.59 19.87 16.27 0.89 0.25 0.41 0.29 0.59 1.15

−1.3 0.6 −0.5 12.3 7.1 32.3 6.5 19.0 27.1 12.6 4.4

6.8

96,136

14,311.9

100

Source: EDB, DOS and computations from NUS Entrepreneurship Centre Survey of Maritime Ancillary Services.

100

12.0

90 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

attributed to improved labour productivity and increasing levels of knowledge intensity in key maritime sectors. Examples of such improvements are evident in the port sector, which is the traditional ‘anchor’ of Singapore’s cluster of maritime-related sectors. In recent years, the Port of Singapore has experienced intense competition from new rival ports in the region, in particular Port Tanjung Pelapas in Malaysia as well as emerging ports in China. These ports enjoy the advantages of operating with substantially lower land and labour costs. Despite this, the Port of Singapore was able to defend its market through a significant increase in productivity achieved via aggressive investment in ICT and automation. As a result of these measures, the Port of Singapore continued to attract significant vessel traffic and retained its position as the top container port in the world and the second-largest port in terms of total cargo throughput. Another segment of the maritime cluster that has upgraded technological capabilities and increased knowledge intensity is the marine engineering industry, comprising the offshore sector and the shipbuilding/repair sector. Until quite recently, the shipbuilding and repair industry was involved in traditional shipyard activities, providing newbuilds and repair services to vessels calling at the Port of Singapore. The industry was able to transform itself by successfully diversifying into major players in the offshore oil and gas construction and marine engineering services sectors. As a result of this, a number of indigenous firms with roots in traditional shipbuilding activities have becoming global leaders in the offshore construction business. More detailed analysis of the offshore sector is provided later in this chapter. Other sectors that achieved high growth through increased productivity and knowledge intensity include bunkering and logistic services. The knowledge-intensive maritime finance, insurance, legal and classification services also registered strong growth, albeit from a small base. 3.3.2.2.2 Role of the state in promoting the development of the international maritime centre in Singapore The Ministry of Transport’s sea transport policy spells out a vision of Singapore as the leading maritime hub in Asia, with a vibrant international maritime centre (IMC) cluster that not only complements and reinforces Singapore’s hub port status, but serves as an additional engine of growth. In 2003, the Maritime and Port Authority (MPA) was appointed as the ‘champion agency’ for the comprehensive development of Singapore from a primarily sea-transport hub towards becoming the leading comprehensive integrated IMC in Asia. MPA plays a leading role in IMC development in the context of a multi-agency coordination approach. As seen in Figure 3.2, the institutional framework for IMC development in Singapore involves a number of ministries and government agencies as well as industry representation through

Cluster Development and Innovation in Singapore

91

Ministry of Transport

Ministry of Education

Ministry of Trade & Industry MPA EDB

Ministry of Finance

Educaton policy logistics oil & gas shipbuilding

MAS Insurance & finance & fiscal policy

Legal services policy

Maritime insurance & finance

IE Singapore

Maritime legal services

ICT policy IDA Cruise

Ministry of Law

Industrial and trade promotion

Industry involvement

Tourism promotion

R&D policy

A*STAR

STB

SMF & Associations: SSA, ASMI etc

Figure 3.2 Overall institutional framework in Singapore for IMC development Source: Wong et al. (2005)

associations and the Singapore Maritime Foundation. The presence of multiple agencies and stakeholders ensures an integrated development approach. Several initiatives are underway to steer the development of Singapore as an IMC. Figure 3.3 details the components of Singapore’s IMC development strategy. MPA is overseeing the expansion of Singapore’s maritime activities from core port and shipping services into bunkering, ship brokering/ chartering, logistic support and classification/surveying services. Another important and related strategic initiative under the IMC thrust is the development of maritime ancillary services such as marine insurance, maritime finance and maritime legal services. MPA has emphasized the attraction of key global players in these fields to Singapore, in addition to promoting local participation. To achieve these goals, the continued vibrancy of the port and shipping services sectors is vital and investments in port upgrading and technological improvement have to continue apace.

92

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Internationalization of port services & shipping

Traditional shipbuilding & repair services

Core port & shipping services industries Port upgrading though investment in R&D and ICT Attract & grow maritime insurance, finance, legal, training services

Diversify into offshore oil & gas & marine engineering services

Listing on Singapore Stock Exchange

Figure 3.3 Singapore’s IMC development strategy

MPA has also worked with the Singapore Economic Development Board and International Enterprises Singapore to expand the traditional shipbuilding and repair industry into offshore oil and gas platform construction and marine engineering services. Diversification in these new areas has created opportunities for R&D and IT projects and provides additional incentives for attracting and growing the maritime ancillary services sectors such as maritime insurance, finance and legal services. 3.3.2.2.3 Government programmes for R&D and innovation in the Singapore maritime cluster The Ministry of Transport has identified the development of Singapore as a maritime R&D centre to be one of the measures for creating a vibrant IMC hub. MPA has instituted a number of initiatives in line with this strategic thrust, among them a Memorandum of Understanding between MPA and the Research Council of Norway (RCN). Signed in 2002, the MOU provides a framework for MPA, RCN and research institutes, academic institutions and industry representatives from Singapore and Norway to collaborate on a number of maritime R&D, Education and Training (RDET) projects that are business- and user-oriented. Areas covered by the MOU include marine environment protection, shipping operations and maritime technology. The scope of the MPA-RCN MOU includes a broad range of activities such as exchange programmes and industrial attachments, education and training courses and cooperation in commercialization results of maritime RDET projects. In addition to the MPA-RCN MOU, MPA has launched and administers the Maritime Innovation and Technology (MINT) Fund. This is a S$100 million fund established to support development programmes under the Maritime Technology Cluster Development Roadmap. $50 million from the Fund has been earmarked for enhancing maritime R&D capabilities.

Cluster Development and Innovation in Singapore

93

Programmes and schemes funded by the MINT fund include: ●

●

●

●

●

TRIDENT Platform (joint programme with EDB): This is a platform for development and test-bedding of maritime innovations called Testbedding, Research and Innovation Development for New-Maritime Technologies (TRIDENT). The programme supports companies and TRIs (tertiary and research institutions) in undertaking maritime-related R&D, innovation development, using Singapore’s port and maritime facilities as test-beds for innovations. Maritime Seed Fund: This fund targets young or growing maritime companies seeking to bring technologies or innovation from concept to commercialization, and established maritime technology companies seeking to embark on further R&D, set up facilities in Singapore or venture overseas. The Fund invests in companies registered in Singapore in exchange for equity based on the three funding levels: up to $50,000 for early-stage startups; up to $300,000 for start-ups; on case-by-case basis for post-start-ups. Joint TRI and MPA R&D Programme: This programme co-funds joint maritimerelated R&D projects by tertiary research institutes (TRIs) and MPA. MPA will participate in the projects as a R&D partner. Projects are to be relevant to the maritime industry and should produce technologies or innovations with the potential to be developed into commercial products, systems or services. TRIs are encouraged to have industry partners in their R&D projects. Maritime Technology Professorships: MPA has set up Maritime Technology Professorships in local universities involved in technological R&D (NUS and NTU). Universities are encouraged to source for industries’ contributions through a dollar-for-dollar matching governmental funding for the industries’ contribution. The aim of the professorships is to encourage R&D relevant to the maritime industry. Maritime Industry Attachment Programme: This programme aims to immerse engineering, IT and science students from TRIs in the maritime industry. From the attachment, maritime R&D concepts and projects can be generated by the students. At the end of their attachment, students can submit their suggestions for maritime R&D projects. Accepted suggestions will be awarded prizes, and final-year or postgraduate R&D projects will be funded by the MINT Fund.

3.3.2.3 Offshore and marine engineering cluster in Singapore The strategy of diversifying traditional shipbuilding and repair into offshore construction and marine engineering services (Figure 3.3) has borne enviable results. Singapore has emerged as a leading offshore and marine engineering cluster in the world, boasting 70 per cent of the global market share for the conversion of floating production storage offloading (FPSO) vessels and 70 per cent of the world market share in jack-up rig construction.

94 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Two indigenous firms, Keppel FELS and SembCorp Marine have emerged as the largest offshore oil rig platform manufacturers in the world. In addition, Singapore has 20 per cent of the world market share for ship repairs and is among the top three global centres for oil and gas (O&G) equipment manufacturing and servicing. In recent years, Singapore has ranked first in Asia by production volumes for six of the top ten players in the oil and gas equipment. Table 3.18 reports the sales revenue of Singapore’s marine and offshore engineering (M&OE) industry in the past ten years. Adopting the Singapore Standard Industrial Classification (SSIC) categories, the M&OE industry is composed of the offshore sectors, shipbuilding sector, ship repair sector and other marine engineering sectors. The two offshore sectors are oil rigs and the other oilfield/gasfield machinery and equipment. In 2006, offshore and marine engineering activities generated over S$12 billion worth of products and services. Over 40 per cent of this amount, or S$5.4 billion, is contributed by the offshore segment of the industry, with another third of output, or S$4.6 billion being generated by the ship repair sector. Revenue in the shipbuilding sector is more modest, at S$1.3 billion, representing around 11 per cent of the industry. In the ten years since 1996, the composition of the M&OE industry has changed. The M&OE industry as a whole has grown at 20.2 per cent per annum between 2001 and 2006, a turnaround from negative growth of 1.8 per cent per annum in the preceding five years. Of the M&OE sectors, the oil rigs sector achieved the highest growth rates, averaging remarkable growth of 49 per cent per annum since 2000, in contrast to annual contraction of 15.8 per cent in 1996–2000. From sales of around S$600 million in the mid1990s, the oil rig sector has multiplied to over S$3 billion in revenues in 2006. The shipbuilding sector also grew at a healthy pace, in the past six years, at 24.3 per cent per annum, outpacing the ship repair sector’s growth of 14.4 per cent per annum. Tracing revenues in the shipbuilding and repair industry over the past 35 years, as shown in Figure 3.4, the most rapid growth has been seen in the recent few years since 2000. Prior to that, growth in all three segments – shipbuilding, ship repair and rig construction – was much more gradual although the ship repair sector experienced a spike in growth in the mid1980s. 3.3.2.3.1 Productivity of offshore sector and shipbuilding/repair sector The growth of the M&OE industry in the past few years is associated with increased productivity, as seen in Table 3.19. Labour productivity as measured by VA per worker has increased from S$42,000 in 1996 to S$55,000 in 2006. The improvement in labour productivity is most pronounced in the oil rig sector, where VA per worker more than tripled from S$22,600 in 1996 to S$76,900 in 2000 and maintaining at S$72,600 in 2006. The other sector

Table 3.18

Sales revenue of marine and offshore engineering industry Offshore

Sales revenue (’000 S$)

Mfg & repair of other oilfield/ gasfield Mfg & repair machinery & of oil rigs equipment

Shipbuilding

Ship repair

Other marine engineering

Building of Repair of ships, tankers & ships, tankers other ocean and other vessels ocean vessels

Building and repair Mfg & repair of pleasure of marine crafts, lighters engine and and boats ship parts

Marine and offshore engineering industry

1,765,819 1,883,759 2,196,496 1,984,042 1,665,994 2,325,768 2,510,931 2,199,417 2,569,369 3,140,156 4,564,192

114,345 99,124 95,887 120,111 141,059 128,887 144,077 158,542 143,863 136,889 200,204

344,613 371,108 322,145 324,915 365,825 449,762 527,371 458,328 559,524 706,509 830,109

4,053,671 4,191,874 4,601,127 4,035,922 3,773,053 4,930,969 5,647,917 5,177,464 6,440,248 8,785,374 12,363,990

−1.44 14.44

5.39 9.21

1.50 13.04

43.56 44.16 36.92

2.82 3.74 1.62

8.50 9.70 6.71

1996 635,110 570,292 623,492 1997 685,180 693,900 458,803 1998 687,907 832,836 465,856 1999 546,618 658,201 402,035 2000 318,922 916,749 364,504 2001 440,369 1,136,767 449,416 2002 977,281 1,045,209 443,048 2003 658,538 1,209,661 492,978 2004 1,061,525 1,320,759 785,208 2005 2,047,241 1,696,337 1,058,242 2006 3,224,299 2,214,282 1,330,904 Growth rate per annum (%) 1996–2000 −15.82 12.60 −12.56 2001–2006 48.91 14.26 24.25 Share in marine & offshore engineering industry (%) 1996 15.67 14.07 15.38 2000 8.45 24.30 9.66 2006 26.08 17.91 10.76 Source: Economic Development Board.

−1.78 20.18 100.00 100.00 100.00

96

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh 9,000 8,000

Shiprepair

Shipbuilding

7,000

Rig construction

Total

S$ million

6,000 5,000 4,000 3,000 2,000 1,000 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

0

Figure 3.4

Shipbuilding and Repair Revenues, 1972–2006

Note: Figures for rig construction sector not available for 1994 and 1995. Source: 1972–1995: Various issues of Singapore Marine and Ocean Engineering Directory/ Singapore Marine Industries Directory, Directory of Singapore Shipbuilding and Offshore Industries 1997; 1996–2006: Economic Development Board.

with marked improvement in labour productivity is the shipbuilding sector. Fixed asset utilization in all the M&OE segments has declined, suggesting that there have been real improvement in efficiency and or gains in total factor productivity to explain the observed growth in labour productivity. Corresponding to improvement in labour productivity, remuneration per worker has improved slightly over the years for the M&OE industry. Average salaries have increased significantly in the offshore and shipbuilding sectors, but declined marginally in the ship repair sectors. On the other hand, value added as a share of revenue has decreased slightly for the M&OE industry as a whole. This was due to declining VA/revenue ratios in the ship repair sector and the oilfield machinery and equipment sector. For the oil rig sector, sectors, the share of VA in revenues generated has increased in the past ten years, while the ratio has remained relatively unchanged in the shipbuilding sector. 3.3.2.3.2 Development of innovation capability in offshore and marine engineering cluster in Singapore Table 3.20 presents a profile of the leading offshore and marine engineering companies in Singapore. There are three major players in the offshore engineering construction sector, namely Keppel FELS, SembCorp Marine and Labroy Marine. All three companies have been in existence for many years, being involved in more traditional shipyard activities in their earlier years. Today, Keppel FELS and SembCorp Marine have consolidated their leading status as the world’s two largest oil rig builders.

Table 3.19

Principal statistics of marine and offshore engineering industry Offshore

Mfg & repair of oil rigs

Mfg & repair of other oilfield/ gasfield machinery & equipment

Shipbuilding

Building of ships, tankers & other ocean vessels

VA/employment (’000 dollars per headcount) 1996 22.6 101.5 70.7 2000 76.9 138.6 81.3 2006 72.6 127.0 129.4 VA/revenue 1996 0.087 2000 0.438 2006 0.142

0.42 0.408 0.288

0.246 0.288 0.240

Ship repair

Repair of ships, tankers and other ocean vessels 32.0 30.4 38.5 0.318 0.322 0.240

Other marine engineering

Building and repair of pleasure crafts, lighters and boats 37.9 47.8 61.3 0.265 0.297 0.300

Mfg & repair of marine engine and ship parts 50.2 44 35.9 0.549 0.514 0.433

Marine and offshore engineering industry 41.9 49.7 55.0 0.291 0.367 0.237

Fixed assets/employment (’000 dollars per headcount) 1996 57.6 70.7 79.4 2000 54.7 67.7 73.8 2006 23.5 69.4 60.4

51.2 37.2 22.7

41.4 34.0 20.8

41.6 25.5 11.2

10.8 10.8 9.3

Remuneration/employment (’000 dollars per headcount) 1996 22.8 52.1 31.4 2000 21.1 57.8 32 2006 37.0 60.3 50.9

23.8 21.8 22.0

28.9 26.9 29.9

37 33.3 25.6

28.1 27.4 29.6

Operating surplus/revenue 1996 20.019 0.155 2000 0.255 0.211 2006 0.077 0.143

0.108 0.142 0.148

Source: Computed from data from Economic Development Board.

0.030 0.059 0.095

0.025 0.096 0.151

0.061 0.079 0.092

0.052 0.129 0.105

98 Poh-Kam Wong, Yuen-Ping Ho and Annette Singh Table 3.20

Leading offshore engineering companies

Company

Sales, S$ million (2005)

Profit S$ million (2005)

Keppel Fels Ltd

1,458.70

127.9

1967 as Far East Shipbuilding Industries Ltd (FESL), involved in rig building

1969 (FESL) 1980 (Keppel)

Sembcorp Marine Ltd

2,119.30

125.6

1963, as Jurong Shipyard, focus on ship repair

1987 (Jurong Shipyard)

515.8

55.9

1978, involved in shipping and shipbuilding

1996

Labroy Marine Ltd

Founding year and activity

Listed

Principal activities today Designs, builds, converts, upgrades and repairs mobile offshore drilling units, floating production systems, production topsides and specialized vehicles. Specializing in a full spectrum of integrated solutions in ship repair, shipbuilding, ship conversion, rig building, topsides fabrication and offshore engineering. Shipbuilding, repair, offshore rig construction and shipping. Entered into the premium rig building market in 2006.

Source: Compiled from company annual reports and S1000.

Although a cluster of Singaporean firms have become major players in offshore oil rig platform production and associated marine engineering services in the world, with production and services contracts around the world, the industry did not engage in significant R&D and patenting activities until recent years. In the offshore rig building sector, capability development was achieved primarily through a continuous process of learning by doing. Firms such as Keppel FELS and Sembcorp Marine initially licensed designs from firms in advanced countries, and with enhanced capabilities, shifted from subcontractor to main contractor roles. Later, these firms acquired rights to design, and subsequently acquired design companies and developed their own inhouse design capabilities. The development of the core rig building firms has stimulated the growth of various marine services firms, as shown in Table 3.21. Such firms are involved in activities such as chartering of offshore supply vessels, logistics

Cluster Development and Innovation in Singapore

99

and procurement support services, equipment distribution and maintenance support. Several of these firms subsequently expanded into oil rig repair, equipment development, IT services and related marine engineering services. 3.3.2.3.3 Development of public R&D institutions and innovation collaboration with offshore industry It is in recent years that the government started to promote R&D activities in the marine engineering field, to support the upgrading of innovation capability among the private sector firms. As part of these efforts, collaborations with institutes of higher education were encouraged. In 2003, the Centre for Offshore Research and Engineering (CORE) was established at the National University of Singapore. CORE was officially launched in 2004 by the Economic Development Board (EDB) and NUS to help strengthen Singapore’s performance as an oil and gas hub in the wake of high growth forecasts for the industry globally. CORE aims to develop advanced technologies and enlarge the talent pool in offshore engineering research through working with other local R&D institutes, international experts and partners in the oil and gas industry. CORE enjoys strong support from A*STAR and collaborates with a number of leading research organizations, including Imperial College London and Det Norske Veritas. Additionally, CORE enjoys strong industry support as firms in the offshore sector recognize the importance of cutting-edge R&D to further maintain competitiveness in this field. Signalling the offshore industry’s support for this initiative, CORE has four founding member companies that have signed a pledge of commitment to the centre’s R&D work. The founding members include the key players in Singapore’s marine engineering sector; Keppel Offshore and Marine, SMOE (a subsidiary of SembCorp), Cooper Cameron and Lloyd’s Register. To date, several R&D projects have been conducted at CORE, including collaborations with local companies as well as overseas organizations such as Norway’s Norsk Hydro in Norway and Delft University of Technology in the Netherlands. The centre has recently, in March 2007, been provided with S$10 million in funding by A*STAR and MPA for a new offshore technology research programme. At the Nanyang Technological University (NTU), a graduate education programme in marine engineering was established in 2004. Jointly developed by NTU and the marine and offshore industry, the programme aims to produce talent to meet the technological and operational demands of the fast-growing offshore industry. At the programme’s inception, six leading offshore and marine engineering companies, including five members of the Association of Singapore Marine Industries (ASMI), pledged over half a million dollars in scholarship to fund students undertaking the course over the next three years.

Table 3.21 Leading offshore support services companies Company

Sales, S$ million (2005)

Profit S$ million (2005)

Founding year and activity

CH Offshore

47.6 (US$ million)

23.0 (US$ million)

Jaya Holdings (inc Jaya Shipbuilding & Engineering and Jaya Offshore )

168.9

Swissco International Ltd *

Swiber

Listed

Principal activities today

1976 as Mico Line Pte Ltd as a offshore service provider

2003

Owns and operates a fleet of anchor-handling tug/supply (AHTS) vessels that provide services to the offshore oil and gas industry.

86.9

1981 as Java Marine Lines, a ship owning company

1994

Fleet-owning and chartering operations and shipyard operations in the building of new vessels. In 2002, made strategic decision to focus on offshore shipping division and reduce conventional shipping division.

13.9

12.4

1970 as Well Industrial and Ship Supply Company, a ship chandler

2004

Marine logistics support services and ship repair and maintenance services to the shipping and offshore oil and gas industries. One of the leading operators of workboats servicing offshore supply vessel services and ‘OutPort-Limit’ (OPL) marine logistics.

25.2 (US$ million)

5.7 (US$ million)

1996, chartering support vessels

2006

An integrated offshore engineering, procurement, construction, installation and commission (‘EPCIC’) contractor with supporting in-house offshore marine capabilities.

KS Energy Services Ltd

269.1

37.1

1974, company dealing in hardware

1999

Distribution of oil & gas products, procurement, engineering and offshore chartering services.

Aqua-Terra Supply Co Ltd**

112.8

6.1

1972, initial focus on product distribution

2004 (Sesdaq) 2006 (Main board)

Service provider and procurement specialist. Material procurement of oil & gas consumables, product distribution for the oil & gas, marine and mining industries.

Ezra Holdings Ltd

72.5

36.1

1992 as Emas Offshore, managing and operating offshore support vessels

2003

Offshore support services – chartering of offshore support vessels, ship management services. Marine services – ship and rig repair, logistics and product sourcing.

Singapore Technologies Marine Ltd

659.8

70.3

1968, design, upgrade and 1990 build commercial vessels, including offshore supply vessels

Turnkey shipbuilding, ship conversion and maintenance & ship repair services, design services.

Note: * Swissco’s associated company, Swiber Holdings Pte Ltd (now operating as Swiber Holdings Ltd), was publicly listed on the Stock Exchange of Singapore Ltd in 2006. Before that Swissco had equity accounting for the results of its associated company Swiber Holdings Pte Ltd (Swiber). ** Subsidiary of KS Energy. Source: Compiled from company annual reports and Singapore 1000.

102

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Another significant move was the establishment of a Marine and Offshore Technology Centre of Innovation (COI (MOT)) at SPRING Singapore, the government agency responsible for enterprise development. The COI (MOT) provides consulting, R&D and technology transfer services to assist SMEs in the marine engineering and equipment industry to upgrade their technological capability. The centre works on a collaborative model and forms partnerships with industry players and, in some cases, with research institutes or centres. Among the services provided by the centre is assistance through the Technology Innovation Programme (TIP). Through the TIP, the centre shares the salary costs of expert manpower and defrays costs of technology innovation projects; up to 50 per cent of allowable costs for technology projects and up to 70 per cent for industry-wide projects. Reflecting the recent policy emphasis on technological development in marine engineering, R&D expenditure in the sector has recorded a significant increase in the late 1990s, as shown in Table 3.22. From annual expenditure of $2.21 million in 1993, R&D spending on marine engineering technologies increased to over $20 million in 2000. In the following six years, the annual amount fluctuated somewhat but averaged at above $20 million annually. Correspondingly, the number of research scientists and engineers (RSEs) involved in marine engineering R&D has increased from 23 to 1993 to 134 in 2006. Similarly, research output in the form of offshore-related patents has increased substantially since the mid-1990s. Since 1978, there have been 23 offshore-related patents granted by the US Patents and Trademarks Office (USPTO) that are either invented by Singaporean residents or assigned to Table 3.22 Key R&D indicators for the Singapore marine engineering sector 1993–2006 R&D expenditure ($ million) 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

2.21 3.39 4.85 13.15 19.09 11.64 17.99 20.95 20.53 124.8 16.27 23.4 19.01 20.66

Source: A*STAR National Survey of R&D in Singapore (various years).

RSEs (FTE) 23 32 41 69 122 87 194 61 82 145.9 77.9 90.4 127.6 133.9

Cluster Development and Innovation in Singapore

103

Singaporean-based interests. Of these 23 patents, 11 have been granted post2000, as shown in Table 3.23. Four of the recently granted patents are technologies developed at two companies, Offshore Technologies and Deepwater Technologies, which are subsidiaries of Keppel FELS. A survey conducted in 2004 covering 13 shipbuilding and offshore marine engineering services firms found four to have close collaboration with public research institutes and universities, and another six reporting sporadic contacts (Wong et al. 2005). Table 3.24 provides some examples of Table 3.23 Offshore patents invented in Singapore or assigned to Singapore interests Total

1978–1989

1990–1999

2000–2007

Foreign assignees ABB Vetco Gray (Texas, USA) FMC Corporation (Illinois, USA) GlobalSantaFe Corporation (Texas, USA) Schlumberger Technology Corporation (Texas, USA) Dril Quip Inc (Texas, USA) A.R.M. Design Development (SG)

7 2 1

1 0 0

4 2 1

2 0 0

1

0

0

1

1

0

1

0

1 1

0 1

0 0

1 0

Singapore assignees Keppel Offshore and Marine subsidiaries: Deepwater Technology Group Pte Ltd (SG) Offshore Technology Development Pte Ltd (SG) Notrans Group (SG) Nortrans Offshore Pte Ltd Nortrans Shipping and Trading Pte Ltd Nortrans Engineering Pte Ltd Prosafe Production Pte Ltd (SG) Individual Assignee (Foster T Manning) (SG) Petroleum Structure Inc (SG) Robin Shipyard Pte Ltd (SG)

16 4

2 0

5 0

9 4

2

0

0

2

2

0

0

2

6 2 3

0 0 0

4 0 3

2 2 0

1 3

0 0

1 0

0 3

1

0

1

0

1 1

1 1

0 0

0 0

Total Singapore invented/ owned offshore patents

23

3

9

11

Note: To identify offshore patents, searched IPC classes E21B, B63B and E02B. The resulting patents were manually vetted for applicability to offshore sector by reading through the full descriptions. Patents applicable to non-offshore vessels are excluded. Source: US Patent and Trademark Office.

104

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Table 3.24 Examples of private and public/IHE collaborations in offshore sector Private sector company

Collaborating institution

Sembawang Marine and Offshore Engineering Pte Ltd

National University of Singapore

Det Norske Veritas

National University of Singapore

Keppel Offshore & Marine

National University of Singapore

Keyser Technologies Pte Ltd

Nanyang Polytechnic (Marine & Offshore Technology Centre of Innovation)

Keppel FELS Ltd

Nanyang Technological University

Nature of collaboration R&D project: ‘Design Automation for Marine/ Offshore Lift Installation of Structures’ funded by NSTB (1996) R&D Project: ‘Lift Dynamics and Decision Support System for Lift Installation of Structures’ funded by NSTB (1998–2001) Joint Industry Project ‘FPSO Fatigue capacity’ (2001–2003) Establishing Keppel Professorship on Ocean, Offshore and Marine Technology in Dept of Civil Engineering (2002). Annual Keppel O&M Lecture. NP-MOTCI was engaged by the client, Keyser, to create a hydro-forming machine for automating expansion joints manufacturing process (2007). Joint R&D for Wave run-up between NTU’s Maritime Research Centre and KFELS

Source: Compiled from information from IHE’s websites and annual reports.

collaborative projects between private sector firms and public research institutes and universities. 3.3.2.4 Case study of leading offshore construction company – Keppel FELS Table 3.25 presents a summary profile of Keppel FELS and its major activities. A profile of SembCorp Marine Limited is also provided as a point of comparison, showing how these two companies have evolved to their status as world leaders in offshore construction. Table 3.26 reports the turnover

Cluster Development and Innovation in Singapore

105

and profits achieved by these two firms in the past 13 years, showing the improved financial figures in the late 1990s resulting from the surge in their contract books in this period. SembCorp Marine operates as a large corporation providing a full suite of services, including offshore construction, and is not organized in such a way that any single corporate entity serves as its offshore division. On the other hand, Keppel FELS is organized as an entity that focuses primarily on offshore activities and, as such, is the basis for this chapter’s case study. Keppel FELS is today the leading designer and builder of sophisticated drilling rigs, having constructed most of the world’s jackups in the past decade. As the wholly owned offshore arm of Keppel Offshore and Marine (Keppel O&M), Keppel FELS’s strategy is also boosted by Keppel O&M’s strategic network of 16 yards worldwide. Keppel FELS has a long corporate history, tracing its beginnings to the incorporation, in 1967, of Far East Shipbuilding Industries Limited (FESL). FESL was primarily involved in rig construction, undertaking a handful of construction projects in its first decade of operations. In 1970, FESL was renamed as Far East Levingston Shipbuilding Limited (FELS), after signing a three-year management agreement with Levingston Shipbuilding Company of Texas, USA. In a separate development, Keppel Shipyard was formed in 1968 to take over the dockyard department of the Port of Singapore. Keppel Shipyard was engaged in the traditional shipyard activities of ship repair and later developed capabilities for shipbuilding. Three years after its formation, Keppel Shipyard acquired a 40 per cent stake in FELS and in subsequent years acquired other shipyards in Singapore and Philippines as part of its expansion plans. In 1980, Keppel Shipyard was listed on the Singapore Stock Exchange and took over complete management of FELS, although the renaming of FELS to Keppel FELS only took place many years later in 1997. Keppel Corporation was incorporated in 1986, with Keppel Shipyard as the major operating division. Since the incorporation of Keppel Corporation, the Keppel group has grown substantially through acquisition of interests in overseas operations as well as incorporation of local and overseas subsidiaries, such as FELS Baltech in Bulgaria. In 2002, Keppel FELS and Keppel Shipyard were integrated to form Keppel Offshore and Marine Limited (KOM), a wholly owned division of Keppel Corporation. With the formation of KOM, Keppel FELS is identified as the offshore engineering arm of KOM while Keppel Shipyard is the shipbuilding and repair division. In addition to Keppel FELS and Keppel Shipyard, a number of other subsidiaries also come under the KOM umbrella. This includes two offshore technology firms – Offshore Technology Development (OTD) and Deepwater Technology Group (DTG) – as well as FELS Offshore Pte Ltd, a holding

Table 3.25

Profile of Keppel FELS and Sembcorp Marine Keppel FELS

Sembcorp Marine

Summary

The wholly owned offshore arm of Keppel Offshore & Marine (Keppel O&M), Keppel FELS has a track record of successful conversions of floating production units and jackup rigs. It also designs, builds, converts, upgrades and repairs the complete range of mobile offshore drilling units, floating production systems, production topsides and specialized vessels.

Sembcorp Marine is a leading global marine engineering group, specializing in a full spectrum of integrated solutions in ship repair, shipbuilding, ship conversion, rig building, topsides fabrication and offshore engineering. The company offers a complete suite of turnkey services to serve the offshore oil and gas industry.

Activities

1967: Incorporation of Far East Shipbuilding Industries Limited – FESL. 1969: Listed on Singapore and Malaysia stock exchanges. 1970: Renamed as Far East Levingston Shipbuilding Limited – FELS. 1978: Formation of Keppel Shipyard. 1971: Keppel Shipyard acquired 40% stake in FELS. 1980: Keppel Shipyard listed on SEX. Keppel Shipyard took over management of FELS. 1997: Renamed FELS as Keppel FELS. 2002: Keppel FELS integrates with Keppel Hitachi Zosen to form Keppel Offshore & Marine Ltd.

1963: Incorporation of Jurong Shipyard. 1968: Incorporation of Sembawang Shipyard. 1987: Jurong Shipyard Ltd is publicly listed. 1988: Jurong Shipyard Ltd acquires Sembawang Shipyard. 2000: Name changed to SembCorp Marine Ltd.

Key acquisitions 1990: 60% equity interest in AMFELS, Texas, USA (1992, AMFELS becomes fully owned subsidiary of FELS). 1994: FELS Baltech (Bulgaria) is incorporated. 1995: 40% stake in Offshore & Marine A.S., Norway. 1999: 77.3% interest in Singapore Petroleum Company (SPC). 2002: Acquires Velrome Botleck, Netherlands, and renames it Keppel Velrome.

2001: Acquisition of 50% of PPL Shipyard (Singapore), 35% of Maua Jurong (Brazil). 2002: Acquisition of 20% of Cosco (Dalian) Shipyard (China); complete acquisition of PT Karimun Sembawang Shipyard (Indonesia). 2003: Acquisition of additional 35% of PPL Shipyard (Singapore). 2004: Acquisition of 30% of Cosco Shipyard Group (China), the enlarged group comprising five shipyards in the key coastal cities of Dalian, Nantong, Shanghai, Zhoushan and Guangzhou.

2005: Acquisition of Sabine Industries Inc. (Texas, USA). 2006: Acquisition of SMOE Pte Ltd, Sembawang Bethlehem Pte Ltd. Evolution

1960s, 1970s: Three projects to fabricate jackup rigs and drillships. 1980s: First contract from an oil company (CONOLCO) positions Keppel FELS as a world-class rig builder. In this period, more than 10 contracts are secured, including first contract for deepwater drilling rig. 1990s: Keppel FELS enters into consortia, alliances and co-ownership arrangements as part of its expansion strategies. In this period, Keppel FELS clinches its first few contracts to fabricate FPSO vessels. In the late 1990s, FELS begins offshore conversion projects. In 1993, Offshore Technology Development (OTD), a wholly owned subsidiary, is set up to develop proprietary technologies in offshore construction. 2000s: Offers integrated total solutions, including new buildings of jackup rigs and mobile rigs, upgrades and conversions of jackups and semisubmersibles, offshore repairs and design and engineering. In the past decade, Keppel FELS has consolidated position as world’s leading designer and builder of jackup rigs, and FPSO and FSO conversions. Niche player in specialized conversions and constructions – including small to medium-sized customized vessels such as Anchor Handling Tug Supply vessels, multipurpose support vessels and cable ships.

Source: Company annual reports, corporate websites.

1960s to mid-1970s: Focus on shipbuilding and ship-repair. 1975 to mid-1990s: Deepened capabilities in shipbuilding and ship-repair. Introduction of ship conversion and offshore activities: jumboization, reefer ship conversion. 1995–2000: Niche shipbuilding and design and construction of large container vessels. Introduce capabilities for offshore conversions: conversion of tankers to FPSO and FSO. Commencement of offshore engineering activities: repair and upgrades of jack-ups and semi-submersibles. 2000s: Proprietary designs of container vessels and design and construction of even larger container vessels. Offers full range of offshore conversions: FPSO, FSO, FPU and specialized FPSO conversions, EPIC FPSO conversion. Introduces rig building service: construction of semi-submersibles and jackups. Offshore production: fixed production platforms and floating production facilities: FPSO, FPU, TLPs, SPARS.

Table 3.26

Turnover and net profit for Keppel O&M Ltd and Sembcorp Marine Ltd, 1993–2005 Sales/turnover (S$ million)

FY 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

Keppel Offshore & Marine Ltd (Keppel FELS figures in brackets) −252.8 −500.9 −654.2 −855 −1,094.7 −969.5 −386.7 −220.5 −350.3 1,889.4 (702.1) 1,441.9 (409.8) 2,393.6 (863.3) 4,068.0 (1,458.7) 5,743.4

Sembcorp Marine Ltd

Net profit (S$ million) Keppel Offshore & Marine Ltd (Keppel FELS figures in brackets)

379.8 334.1 325.4 357.2 665.1 933.7 921 763 854.5 1,011.5 1,068 1,362.8 2,119.3 3,545.1

Note: Keppel Offshore and Marine Ltd incorporated in 2002, no financial figures prior to 2002. Source: Singapore 1000 (various years); Company annual reports.

−27.3 −47.1 −50.4 −54.6 −21 −20.2 −54.7 −85.3 −111.4 202.9 (101.7) 109.6 (100.0) 191.9 (99.8) 228.4 (128.0) 458.8

Sembcorp Marine Ltd 68.8 52 39 35.5 47.6 72.9 76.8 75.3 80.9 93.2 78.2 98 125.6 228.2

Cluster Development and Innovation in Singapore

109

company for managing the overseas subsidiaries involved in offshore related activities. It is instructive to analyse the evolution of KFELS’ activities and strategies over the years to understand how it has achieved its current market leader status in the construction of offshore drilling rigs. From its inception in 1967 throughout the 1970s, FELS managed a small number of projects, delivering four drillships and rigs. In the 1980s, after coming under the management of Keppel Shipyard and having access to the Keppel’s substantial yard facilities, FELS embarked on a number of joint ventures to broaden its markets and scope of services. These include ventures with Finland’s KONE corporations and Nanhai Oil Equipment Repairs and Maintenance Company Shenzhen. During the 1980s FELS secured its first contract from an oil company, CONOLCO and successful delivery of the Tension Leg Wellhead Platform to CONOLCO positioned FELS as a world-class rig builder. There was a flurry of activities with more than ten contracts secured, including the first contract for a deepwater drilling rig. The 1980s saw FELS venturing into technological development. In 1981, CAD/CAM technology was introduced to FELS, marking a technological upgrade in its design operations. A subsidiary firm, Offshore CIM Engineering Projects (OECP) was later formed in 1990 to further develop computer-integrated manufacturing technology. In 1985, FELS launched into its first rig R&D initiative in an agreement with Foramer SA, Friede and Goldman to build and own the Columbus Explorer, a MOD V jackup. However, under this arrangement, the technology and rig design resided with Friede and Goldman. In the 1990s, Keppel FELS aggressively entered into consortia, alliances and co-ownership arrangements as part of its expansion strategies. Significant acquisitions included stakes in AMFELS in the US and Offshore and Marine A.S. in Norway. In the early 1990s, FELS clinched its first few contracts to fabricate FPSO vessels and in the late 1990s began working on offshore conversion projects. Several key technological developments took place in this period, allowing for FELS to deepen their capabilities and embark on FPSO construction and conversion projects. The Keppel Group embarked on a series of scholarship and study schemes, sending Keppel staff to be trained in technologically advanced countries such as the UK, USA, Germany, France, Norway and Japan. Through these programmes, Keppel brought technology into Singapore and laid the foundations for Singapore to develop its own infrastructure for offshore and marine technology. In 1993, Offshore Technology Development Pte Ltd (OTD), a wholly owned subsidiary of KOM, was set up with the support of the Singapore National Science and Technology Board (NSTB) to develop proprietary technologies in offshore construction. OTD made great strides in a relatively short time and delivered its first proprietary design in 2000. At the same time, Keppel FELS continued to acquire technology

110

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

from external sources. In 1997, it acquired the rights to the Friede and Goldman MOD V and MOD VI rig designs. These developments indicated the dual mode being adopted for technological development at Keppel FELS: acquisition and enhancement of external technology while developing in-house capabilities. In the late 1990s and into the 2000s, Keppel FELS has experienced the most productive period in its history. Contracts are being secured at a rapid rate and construction activities have grown apace to keep up with market demands. As a result, Keppel FELS has consolidated its position as the world’s leading designer and builder of jackups, FPSO and FSO conversions. At the same time, it has extended its offerings to provide integrated total solutions, not only in new buildings but also upgrades and conversions, offshore repairs and design and engineering solutions. A KOM subsidiary, Keppel Singmarine, has also developed capabilities to become a niche player in conversions and construction of specialized offshore vessels, such as anchor handling tug supply vessels and cable ships. A significant development at Keppel FELS and KOM, more generally, is the increase in emphasis and resources devoted to R&D. In 2004, the Keppel Technology Advisory Panel was set up, comprising eminent scientists, industrialists and practitioners in the offshore field. In line with management’s R&D thrust, the KOM Technology Centre (KOMTECH) was inaugurated in 2007 to step up the growth of in-house competencies for R&D and technological development. Continued investment in R&D has borne fruit, as seen by KOM’s success in creating IP assets in rig designs and systems. The two technology firms under the KOM umbrella – OTD and DTG – had two patents each granted by the US Patent Office between 2000 and 2007. At present, OTD has another six published applications with the USPTO. OTD’s patented rig designs and equipment feature extensively on rigs constructed by Keppel FELS while DTG’s patents on semisubmersible vessels are suitable for operations worldwide. As of December 2007, it was estimated that KOM’s proprietary technologies have generated S$15 billion of contracts, including projects currently under construction.3 KOM sees itself as a main driver in developing Singapore as a centre of excellence for offshore and marine technology. When KOM was inaugurated in 2002, the Executive Chairman of Keppel Corporation kick-started Keppel’s Centre of Excellence. The centre aimed to conduct extensive R&D activities and to build a repository of expertise, experience and knowledge. One of the key initiatives was the establishment of the Keppel Professorship in Ocean, Offshore and Marine Technology at the National University of Singapore (NUS). KOM sponsors the Keppel Professorship with an initial funding of S$1.5 million to initiate research projects and product and technology development. In addition to the Keppel Professorship, KOM has worked with NUS to launch a series of Keppel Offshore and Marine Lectures,

Cluster Development and Innovation in Singapore

111

attracting audiences from industry as well as the student and research communities. KOM also sponsors visiting professors in related fields and supports international conferences such as the International Conference on Technology and Operation of Offshore Support Vessel (OSV Singapore 2005) in September 2005. KOM is also the First Founding Member of the Centre for Offshore Research and Engineering (CORE) at the National University of Singapore.4 CORE was unveiled in 2004 by the Economic Development Board and EDB in an official launch, during which the founding members signed a ‘Pledge of Commitment’ to reinforce their support for the R&D initiatives of CORE. KOM has since gone on to work with CORE on a number of collaborative R&D projects. A KOM-CORE collaboration on jackup platforms and spudcan foundation has resulted in the filing of a patent. Aside from its collaborations with NUS, KOM has also forged links with the Nanyang Technological University (NTU). In 2005, KOM endowed $50,000 into two student schemes at NTU’s School of Mechanical and Aerospace Engineering. The two endowment schemes, a medal and book prize, will benefit final year students excelling in selected subjects. The endowment is viewed by KOM’s COO, Mr Tong Chong Heong, as an investment in the new generation of talent needed in the maritime industry. KOM is also one of the seven sponsors of the Marine and Offshore Engineering specialization and scholarship programme introduced at NTU in 2004.

3.4 Conclusion 3.4.1 Common elements in cluster development strategies Several common elements were observed between strategies for developing the biomedical sciences cluster and those for the offshore marine engineering cluster. In both cases, there was significant involvement from the state. The government of Singapore adopted a top-down, coordinated multiagency approach to developing the two clusters, comprising measures for investment promotion, promoting R&D through developing public R&D institutes and providing incentives for private sector R&D, and infrastructure and manpower development. Another key cluster development strategy is the attraction of a critical mass of anchor firms or institutions to jump-start the cluster. For the BMS cluster, foreign pharmaceutical MNCs were incentivized to establish operations for knowledge-intensive services such as R&D. In the offshore engineering cluster, firms in the traditional shipbuilding and repair sector were diversified into offshore construction activities. Attracting anchor firms has involved investment promotion policies that are proactive, targeted and concentrated. Another strategy for building critical mass is geographic

112

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

agglomeration of cluster players, such as that achieved by the Biopolis facility. With the presence of anchor firms in a newly formed cluster, new entrants are able to leverage on the expertise of the early entrants for learning and knowledge transfer, thus facilitating cluster growth. In Singapore, government involvement in cluster development is clearly seen in the investments made in cluster-wide ecosystem development. For both biomedical and offshore clusters, relevant government agencies have made proactive investments to develop core infrastructure (e.g., setting up R&D institutions) and skills (e.g., working with educational institutions to introduce courses and programmes). Additionally, the government has successfully leveraged on anchor firms or institutions to stimulate development of more specialized resources, supporting industries and services. For instance, the well-developed health-care sector provided a basis for developing medical technology services and R&D activities and stimulated the formation of start-up firms in drug discovery and medical devices. In the offshore sector, the activities of Keppel FELS and Sembcorp Marine created opportunities for the offshore support service sector. 3.4.2

Differentiating elements in cluster development strategies

The timing of cluster development policies has differed in the two clusters in question. The state was involved in the conception of the BMS cluster, playing a leading role in its creation and growth, beginning with the BMS Hub initiative launched in June 2000 and followed by a series of broadbased policy initiatives. In the marine engineering services cluster, the role of the state is more supportive and state involvement took place when the original shipbuilding and repair cluster (from which the offshore engineering cluster was developed) was at a mature stage. The private sector companies took the lead in growing the offshore cluster and have been doing so since the 1990s. There was a notable difference in the role played by indigenous firms versus the role of foreign MNCs in the two clusters. While policies to attract foreign global MNCs were critical in the development of pharmaceutical manufacturing, policies to nurture indigenous firms were more important in the case of marine engineering services. For both clusters, R&D and innovation were important elements of their development paths. However, R&D and innovation played different roles at different stages of cluster development. For BMS, the establishment of R&D capabilities, generation of IP and their subsequent commercialization were critical right from the start in the cluster development process. On the other hand, for marine engineering, cluster development started with manufacturing, learning by doing, and gradual accumulation of tacit process knowledge and innovation capabilities. The role of public R&D institutions came later on, and the creation of IP was not important until much later in the cluster development process.

Cluster Development and Innovation in Singapore

113

The contrasting nature of BMS and offshore engineering meant that there would be different emphasis on the local or regional production base. Although Singapore was able to attract a sizable manufacturing base of global pharmaceutical MNCs, this was not sufficient to jump-start entry into the biotech/life science industries. Because early stage start-ups are the key drivers for BMS, venture capital was essential. Rather than foreign investments, the attraction of global talent was more critical to build the critical mass of expertise needed. The situation is different in the offshore engineering cluster where leveraging on the local production base was more effective in stimulating cluster growth. The prior establishment of indigenous manufacturing in shipbuilding/repair played a significant role in facilitating the learning of know-how and development of innovation capabilities in marine engineering. Similarly, the role of local/regional markets was different for the two clusters. The initial creation of the BMS cluster was driven primarily by perceived global market growth opportunities rather than local/regional market opportunities; nevertheless, in the future, a greater focus on leveraging local/regional health-care markets by engaging local/regional hospitals in translational research will be important in sustaining cluster development. In contrast, the presence of offshore oil and gas production in Southeast Asia (particularly Indonesia) provided an early opportunity for Singaporean firms to learn and develop capabilities in marine engineering services. Subsequent development of the cluster was driven by globalization of these local firms, while growth opportunities from the regional markets continue to be pursued. For the BMS cluster, establishing a strong IP protection regime and providing a strong public policy framework regarding bioethics and standards for clinical trials were critical to attracting foreign interest to participate in the cluster. In contrast, for the marine engineering services cluster, tacit process knowledge is more important, although the role of the state in providing transparent rule of law and reputation for trust and security also contributed. Apart from IP policies, supporting policies also played vastly different roles in the two clusters. The role of the state in developing a centralized physical infrastructure (Biopolis) was critical to the development of the BMS hub. In the case of marine engineering, the role of the state was in making land available for shipbuilding/repair and developing and maintaining a world-class maritime port infrastructure that facilitated maritime activities in general. In terms of policies related to finance, the active promotion of Singapore as a financial hub for maritime industries, in particular the listing of stocks of maritime companies on the Singapore Stock Exchange, facilitated the development of indigenous firms in the marine engineering services industry in terms of fund-raising. In contrast, the Singapore stock market has

114

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

played little role in attracting biotech firms, although the role of the state in providing venture capital (BioOne) has been critical in attracting some biotech firms. 3.4.3 Implications for other economies The case studies of the BMS and offshore clusters in Singapore show that it is possible to accelerate the development of knowledge-based industrial clusters through public policy. Singapore’s experience suggests that this is likely to require a coordinated, strategic approach involving multiple government agencies and sustained investment over a long period. That being said, there is no ‘one-size-fits-all’ approach for the role of the state in cluster development. While all the key component elements of a knowledge cluster need to be developed, the specific roles and timing of state involvement depend on the nature of the industrial clusters to be developed. As Singapore’s experience in the BMS and offshore clusters has shown, there are multiple factors to consider, such the nature of technologies or processes involved and the market environment. Another lesson to be learnt from the Singapore experience is the potential trade-off between sourcing for external capabilities vs internal expertise. Because of Singapore’s small population, the strategy to leverage on global MNCs and foreign talents has long been in practice. This can be effective in accelerating the development of a cluster, as seen in the BMS industry. However, this may also run the risk of slowing indigenous capability development.

Notes 1. Interview with Kong Hwai Loong of the Biomedical Research Council. 2. The Straits Times. Chang Ai Lien. ‘Maintaining S’pore’s lead in stem-cell race’, 5 September 2001. 3. Estimate mentioned in speech by Mr George Yeo, Minister for Foreign Affairs, at Keppel Offshore and Marine’s fifth anniversary celebrations. Available as a media release from Government of Singapore: http://app.sprinter.gov.sg/data/ pr/20071203983.htm (accessed 28 January 2008). 4. Information on the establishment of CORE is drawn from Cheah et al. (2006).

References Agency for Science, Technology and Research (A*STAR) (various years) ‘National Survey of R&D in Singapore,’ Singapore: Agency for Science, Technology and Research (previously National Science and Technology Board). —— (2007a) ‘Inaugural Translational and Clinical Research Flagship Programme Awarded to Gastric Cancer Research,’ press release dated 4 July 2007, www.a-star. edu.sg/astar/attach/press_release/IAC_Press_Release.pdf (accessed January 2008).

Cluster Development and Innovation in Singapore

115

—— (2007b) ‘Phase 2 of Singapore’s Biomedical Sciences Initiative Gains Momentum with Neuroscience Partnership and New Infrastructure for Clinical Research,’ press release dated 19 October 2007, www.a-star.edu.sg/astar/attach/press_release/IAC_ Press_Release.pdf (accessed January 2008). A*STAR website: www.a-star.edu.sg/ (accessed January 2008). Biomed-Singapore (2003), www.biomed-singapore.com/bms/gi_mc.jsp (accessed January 2008). Biomedical Research Council (BMRC) (2007) ‘Exceptional Growth for Singapore’s Biomedical Sciences Industry’, www.biomed-singapore.com/bms/sg/en_uk/index/ newsroom/pressrelease/year_2007/6_feb_-_exceptional.html (accessed January 2008). Cheah, Kok-Keong, Chun-Fai Leung and Yoo-Sang Choo (2006) ‘Collaboration with Research Centres to Promote CEE and Industry Outreach,’ in Proceedings of the 10th IACEE World Conference on Continuing Engineering Education, 2006, Mervyn Jones, Andreas Krieger, Franz Reichl, and Andreas Steiner Ed., Vienna: Vienna University of Technology. DP Information Network Pte Ltd (various years) Singapore 1000. Singapore: DP Information Network Pte Ltd. Economic Development Board (EDB) (various years) Report on the Census of Industrial Production. Singapore: EDB. —— (various years) Report on the Census of Manufacturing Activities. Singapore: EDB. Finegold, D., P. K. Wong and T. C. Cheah (2004) ‘Adapting a Foreign Direct Investment Strategy to the Knowledge Economy: The Case of Singapore’s Emerging Biotechnology Cluster.’ European Planning Studies, Vol. 12, No. 7, pp. 921–941. International Institute for Management Development (IMD) (2006) The World Competitiveness Yearbook 2006. Lausanne, Switzerland: IMD. Kuchiki, A. (2005) ‘A Flowchart Approach,’ in Industrial Clusters in Asia. Analyses of their Competition and Cooperation, A. Kuchiki and M. Tsuji Eds, New York: Palgrave Macmillan. National Research Foundation (NRF) (2007) NRF CREATE webpage: www.nrf.gov.sg/ NRF/otherProgrammes.aspx?id=366 (accessed January 208). USPTO database, www.uspto.gov (accessed January 208). Wong, P. K. (2001) ‘Leveraging Multinational Corporations, Fostering Technopreneurship: The Changing Role of S&T Policy in Singapore.’ International Journal of Technology Management, Vol. 22, No. 5/6, pp. 539–567. —— (2002) ‘Globalization of American, European and Japanese Production Networks and the Growth of Singapore’s Electronics Industry.’ International Journal of Technology Management, Vol. 24, No. 7/8, pp. 843–869. —— (2007) ‘Commercializing Biomedical Science in a Rapidly Changing “Triplehelix” Nexus: The Experience of the National University of Singapore.’ Journal of Technology Transfer, Vol. 32, pp. 367–395. Wong, P. K., Y. P. Ho, and K. K. Ng (2004) ‘Study of Singapore’s Maritime Cluster. Phase 1 Report: Singapore Findings,’ Report submitted to the Maritime and Port Authority of Singapore. —— (2005) ‘Study of Singapore’s Maritime Cluster. Phase 2 Report: Policy Review and Recommendations,’ Report submitted to the Maritime and Port Authority of Singapore.

116

Poh-Kam Wong, Yuen-Ping Ho and Annette Singh

Wong, P. K., Y. P. Ho, and A. Singh (2007) ‘Performance Indicators for Singapore’s International Maritime Centre,’ Report submitted to the Maritime and Port Authority of Singapore. Zucker, L. and M. Darby (1996) ‘Star Scientists and Institutional Transformation: Patterns of Invention and Innovation in the Formation of the Biotechnology Industry.’ Proceedings of the National Academy of Sciences, USA, 93, pp. 12709–12716.

4 Empirical Analysis of the Relationship between Upgrading and Innovation of Japanese SMEs and Industrial Clustering Shoichi Miyahara and Masatsugu Tsuji

4.1

Introduction

The Japanese economy has been suffering from a long recession since the early 1990s. Since then, countless measures to revitalize the industrial sector have been implemented by all levels of government, from central to local, and a significant amount of public funding has been poured into various projects, such as promoting venture businesses or supporting academia/industry/government collaboration. The reality of the Japanese economy, however, shows that revitalization has not occurred. Thus far, such policy measures have not been successful in promoting Japan’s revitalization; moreover, the gap in economic circumstances between metropolitan and rural areas, and between large companies and SMEs (small and medium-sized enterprises) has been enlarging. It is recognized that the revitalization of regional industries is one way to cope with these issues. In so doing, focus has been placed on upgrading regional industries and SMEs to equip them with higher technology and management. One means to achieving this is the industrial cluster policy, which aims to revitalize regional industries and SMEs by agglomerating firms which are large or new start-ups, research institutions related to high or low technologies, and universities with research of cutting-edge technology. The rationale is provided by Fujita et al. (1999), Krugman (1991), Porter (1980) and Saxenian (1994), for instance. The essence of these theories, in the present context, lies in the flow of information generated by agglomeration; that is, in regions where firms and research institutions cluster, collaboration and competition among those parties and organizations create not chaos, but rather the ‘coherent power’ of vitalization. We refer to this process as the ‘endogenous innovation process’.1 Once a region develops sufficient power 117

118

Shoichi Miyahara and Masatsugu Tsuji

to create something new, the process can repeat itself to yield another such upgrading and innovation. The authors have been conducting research so far in order to formulate how industrial clustering occurs mainly in East Asian economies, and the hypothesis we are postulating is referred to as the ‘Flowchart Approach’, initiated by Kuchiki (2007). Based on accumulated studies such as Kuchiki and Tsuji (2005), Tsuji et al. (2006), Tsuji et al. (2007) and Tsuji et al. (2008), the Flowchart Approach has been verifying and elaborating. Industrial clustering itself, however, is not the final aim to vitalize the regional as well as national economies, but it is one effective method to trigger economic activities. One more important role of agglomeration is that it is fundamental basis of innovation or industrial upgrading in industrial clusters. This role of clustering has been emphasized by many authors, such as Porter (1980), Saxenian (1994), and Fujita et al. (1999), as already mentioned. This chapter thus aims to initiate the so-called ‘Flowchart Approach to endogenous innovation process’ inside an industrial cluster, and makes an attempt to postulate how industrial clustering transforms into the upgrading and innovation process. In order to analyse this process, at first we have to clarify how firms inside a cluster are conducting innovation and upgrading and how their activities are different from those outside a cluster. In so doing, this chapter aims to verify the hypothesis that a relationship exists between innovation and industrial clustering formed by regional SMEs. We conducted an extensive mail survey to 5,000 SMEs which were authorized as ‘innovative’ by the Small and Medium Enterprise Agency, and divided these 5,000 SMEs into two groups, those inside or outside a cluster. By comparing the two groups, we analyse how industrial clusters and regional research institutions influence innovations and the upgrading of SMEs. This is thought to be a preliminary step to postulate the endogenous innovation process.2 The chapter consists of the following sections: Section 4.2 presents the contents of the mail survey conducted in October and November 2007; in Section 4.3, the methodology of the statistical analysis and two models, namely the upgrading and innovation models, are explained; and the results of estimations using the upgrading and innovation models are presented in Sections 4.4 and 4.5 respectively. In the final section, conclusions and suggestions for further research will be briefly presented.

4.2

Results of the mail survey

First, the contents of the mail survey, conducted in October and November 2007, and a summary of the results are presented. 4.2.1 Objectives of the mail survey The objective of this mail survey was to obtain and analyse data to verify two hypotheses: (1) the relationship between SME upgrading and innovation and

Relationship between Innovation of SMEs and Clustering

119

industrial clusters; and (2) the relationship between upgrading and innovation, and regional collaboration with research facilities such as universities and other public research institutions. To verify these, we selected SMEs authorized as ‘innovative’ by the Small and Medium Enterprise Agency, which aims to support SMEs and assist their survival in the current severe circumstances. The Agency authorizes SMEs as innovative and supports the restructuring of their businesses to expand into new fields or the upgrading of their technologies.3 In this chapter, we divide them into two groups: that is, SMEs inside and outside a cluster.4 We then compare these two groups in order to examine whether there are differences in upgrading and innovation; that is, we examine how industrial clusters and regional collaboration promote the upgrading and innovation of SMEs. 4.2.2

Characteristics of the respondent SMEs

We chose the SME sample as follows: we calculated the share of each prefecture with regard to the total number of authorized SMEs, and multiplied this by 5,000, which is the total number of mail questionnaires we wanted to send. This results in the number of mails to be sent to each prefecture. Then, we divided this number by the number of years in which there are SMEs authorized according to share, and thereby obtained the number of firms to choose in each prefecture. Finally, we selected SMEs inside a cluster and those outside a cluster. The questionnaire was then sent in November 2007 to 2,000 SMEs inside and 3,000 outside a cluster. A total of 889 valid responses were received. The overall response rate was 17.8 per cent. The numbers of replies from SMEs inside and outside a cluster are 316 (35.6 per cent) and 573 (64.5 per cent) respectively. In order to analyse the role of regional research facilities which collaborate with SMEs on new projects, the distance between SMEs and collaborating partners is crucial, and accordingly we asked about this in Question VII 6–6, namely, ‘How many minutes does it take by car (or equivalent) to reach partners (other companies, universities, and research institutes) with whom you collaborate on new projects?’ The replies to this question are summarized in Table 4.1. The interesting results are: (1) more than half the SMEs are located within 1 hour’s drive from collaborating partners; and (2) the percentage of SMEs located within this distance are greater than that of SMEs outside a cluster. This indicates that in industrial clusters, SMEs and collaborating research facilities are located closely each other. Table 4.2 indicates the distribution of the year of establishment, which is evenly distributed, particularly in total, except over 50 years. However, SMEs outside a cluster have rather large variance. Tables 4.3 and 4.4 show the size of SMEs in terms of capital and employees respectively. The numbers of SMEs, which firm sizes in terms of capital are 10–20 million yen and over 50 million yen, account for more than 50 per cent. Table 4.5 shows industry; most of the SMEs are engaged in the manufacturing sector, and this bias is often found in the data related to the Small and

120

Shoichi Miyahara and Masatsugu Tsuji

Table 4.1 Distance between SMEs and collaborating partners Inside of cluster

Within 30 minutes 30 minutes to 1 hour 1 to 1.5 hours 1.5 to 2 hours 2 hours Too far to go by car No reply Total

Outside of cluster

Total

Freq.

%

Freq.

%

Freq.

%

92

29.11

129

22.51

221

24.86

87

27.53

139

24.26

226

25.42

10 16 24 9

3.16 5.06 7.59 2.85

33 46 56 15

5.76 8.03 9.77 2.62

43 62 80 24

4.84 6.97 9.00 2.70

78 316

24.68 100.00

155 573

27.05 100.00

233 889

26.21 100.00

Source: Authors.

Table 4.2 Year of establishment Inside of cluster Freq. 0–10 years ago 10–20 years ago 20–30 years ago 30–40 years ago 40–50 years ago Over 50 years ago No reply Total

38 44 52 51 49 77 5 316

% 12.03 13.92 16.46 16.14 15.51 24.37 1.58 100.00

Outside of cluster

Total

Freq.

%

Freq.

%

64 114 80 101 57 136 21 573

11.17 19.90 13.96 17.63 9.95 23.73 3.66 100.00

102 158 132 152 106 213 26 889

11.47 17.77 14.85 17.10 11.92 23.96 2.92 100.00

Source: Authors.

Medium Enterprise Agency. Table 4.6 explains the specific category within manufacturing, showing that food, metal, general machinery, and electrics are the major industries. Table 4.7 indicates relationships with other firms through sales: most SMEs were independent manufacturers or selling their products to non-keiretsu firms.5 The total amount of sales in the most recent year is shown in Table 4.8. Most of the SMEs have a larger amount of sales; in particular, those in the 100–300 million yen and over 1 billion yen categories accounted for a greater than 25 per cent share, respectively. Table 4.9 indicates the trend in sales within the most recent three years, showing that more than half have been increasing sales, but that the share of SMEs inside a cluster with increasing or decreasing sales is larger than that outside of SMEs outside

Relationship between Innovation of SMEs and Clustering

121

Table 4.3 Amount of capital Inside of cluster

Outside of cluster

Million yen

Freq.

Freq.

Under 10 10–20 20–30 30–40 40–50 Over 50 0 No reply Total

32 136 49 46 0 49 1 3 316

% 10.13 43.04 15.51 14.56 0.00 15.51 0.32 0.95 100.00

74 222 81 58 0 130 2 6 573

Total

%

Freq.

%

12.91 38.74 14.14 10.12 0.00 22.69 0.35 1.05 100.00

106 358 130 104 0 179 3 9 889

11.92 40.27 14.62 11.70 0.00 20.13 0.34 1.01 100.00

Source: Authors.

Table 4.4 Number of employment Inside of cluster Freq. Under 4 4–9 10–19 20–49 50–99 Over 100 No reply Total

25 57 66 101 42 23 2 316

Outside of cluster

%

Freq.

7.91 18.04 20.89 31.96 13.29 7.28 0.63 100.00

42 98 126 150 107 47 3 573

% 7.33 17.10 21.99 26.18 18.67 8.20 0.52 100.00

Total Freq.

%

67 155 192 251 149 70 5 889

7.54 17.44 21.60 28.23 16.76 7.87 0.56 100.00

Source: Authors.

a cluster. In addition, related to profits, shown in Table 4.10, a greater percentage of SMEs inside a cluster achieved a surplus than SMEs outside a cluster. From these observations, the business performance of authorized SMEs inside clusters is, in general, better than that outside clusters. In this regard, The New Business Promotion Act for SMEs is considered successful. Regarding R&D expenditures, Table 4.11 indicates that nearly 50 per cent of SMEs spent less than 5 per cent of total sales on R&D, but that nearly 12 per cent did not invest in R&D. These are because of the small sizes of their firms. There are no large differences between SMEs inside and outside a cluster, but the latter seemed to have greater R&D expenditures. Finally, distribution by year of authorization by the Small and Medium Enterprise Agency is indicated in Table 4.12. Some SMEs were authorized

122

Shoichi Miyahara and Masatsugu Tsuji

Table 4.5 Category of industry Inside of cluster

Construction Manufacturing Wholesale/retail Information and communications Traffic Other service industry Others No reply Total

Outside of cluster

Total

Freq.

%

Freq.

%

Freq.

%

17 231 32 5

5.38 73.10 10.13 1.58

34 420 43 15

5.93 73.30 7.50 2.62

51 651 75 20

5.74 73.23 8.44 2.25

2 14 14 2 316

0.63 4.43 4.43 0.63 100.00

7 44 26 2 573

1.22 7.68 4.54 0.35 100.00

9 58 40 4 889

1.01 6.52 4.50 0.45 100.00

Source: Authors.

Table 4.6 Category of manufacturing

Food Textiles Wood Print Chemistry Plastic Rubber Leather Steel Metal General machinery Communication Electric Transport Precision equipment Others No reply Total Source: Authors.

Inside of cluster

Outside of cluster

Freq.

%

Freq.

%

Freq.

%

6.49 5.19 0.87 6.06 2.60 3.90 1.30 0.00 2.60 19.91 9.96 3.90 8.66 4.33 6.93 17.32 0.00 100.00

65 15 21 18 9 20 2 0 8 49 53 19 31 16 25 65 4 420

15.48 3.57 5.00 4.29 2.14 4.76 0.48 0.00 1.90 11.67 12.62 4.52 7.38 3.81 5.95 15.48 0.95 100.00

80 27 23 32 15 29 5 0 14 95 76 28 51 26 41 105 4 751

10.65 3.60 3.06 4.26 2.00 3.86 0.67 0.00 1.86 12.65 10.12 3.73 6.79 3.46 5.46 13.98 0.53 100.00

15 12 2 14 6 9 3 0 6 46 23 9 20 10 16 40 0 231

Total

Relationship between Innovation of SMEs and Clustering

123

Table 4.7 Subcontracting Inside of cluster

Outside of cluster

Freq.

%

Freq.

%

Freq.

200 25

63.29 7.91

367 80

64.05 13.96

567 105

63.78 11.81

85

26.90

132

23.04

217

24.41

14 8 316

4.43 2.53 100.00

23 12 573

4.01 2.09 100.00

37 20 889

4.16 2.25 100.00

Self-products Orders from keiretsu company Orders from non-keiretsu company Others No reply Total

Total %

Source: Authors.

Table 4.8 Recent annual sales

Under 50 million yen 50–100 million yen 100–300 million yen 300–500 million yen 500 million–1 billion yen Over 1 billion yen No reply Total

Inside of cluster

Outside of cluster

Freq.

%

Freq.

25 31 75 44 58 82 1 316

7.91 9.81 23.73 13.92 18.35 25.95 0.32 100.00

56 56 152 69 103 136 1 573

% 9.77 9.77 26.53 12.04 17.98 23.73 0.17 100.00

Total Freq.

%

81 87 227 113 161 218 2 889

9.11 9.79 25.53 12.71 18.11 24.52 0.22 100.00

Source: Authors.

Table 4.9 Trend of sales amount within recent 3 years

Decreasing Same Increasing No reply Total Source: Authors.

Inside of cluster

Outside of cluster

Total

Freq.

%

Freq.

%

Freq.

%

56 93 166 1 316

17.72 29.43 52.53 0.32 100.00

94 183 294 2 573

16.40 31.94 51.31 0.35 100.00

150 276 460 3 889

16.87 31.05 51.74 0.34 100.00

124

Shoichi Miyahara and Masatsugu Tsuji

Table 4.10 Balance of revenues and costs in recent 3 years

Surplus Balanced Deficit No reply Total

Inside of cluster

Outside of cluster

Total

Freq.

%

Freq.

%

Freq.

%

159 108 46 3 316

50.32 34.18 14.56 0.95 100.00

275 196 98 4 573

47.99 34.21 17.10 0.70 100.00

434 304 144 7 889

48.82 34.20 16.20 0.79 100.00

Source: Authors.

Table 4.11 Ratio of R&D expenditures to total sales Inside of cluster

Outside of cluster

Total

%

Freq.

%

Freq.

%

Freq.

%

Under 5 5–10 10–20 Over 20 0 No reply Total

125 43 31 22 30 65 316

54.11 18.61 13.42 9.52 12.99 28.14 100.00

219 83 57 20 57 137 573

52.14 19.76 13.57 4.76 13.57 32.62 100.00

344 126 88 42 87 202 889

45.81 16.78 11.72 5.59 11.58 26.90 100.00

Source: Authors.

more than once, and the number of authorizations has been increasing, except during 2007. 4.2.3 Upgrading and innovation From replies of respondents, characteristics of upgrading and innovations can be seen as follows. 4.2.3.1 Upgrading In this chapter, we define industrial upgrading as represented by the following examples: (1) from being subcontractors for simple work to producing intermediate goods; (2) from producing intermediate goods to final products; and (3) from simple to complex or precision work. Question V consists of the following six sub-questions regarding upgrading and innovation: V.1. We upgraded business activities; for example, we upgraded from being subcontractors for simple work to producing intermediate goods, from producing intermediate goods to final products, or from simple to complex or precision work.

Relationship between Innovation of SMEs and Clustering

Table 4.12

Year of authorization Inside of cluster

1990 1991 1995 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 No reply Total

125

1st

2nd

0 0 0 1 2 9 26 13 14 35 38 57 85 18 18 316

0 0 0 0 0 0 0 1 1 4 5 6 22 11 266 316

3rd 0 0 0 0 0 0 0 0 0 0 1 0 0 1 314 316

Outside of cluster

4th

1st

2nd

3rd

0 0 0 0 0 0 0 0 0 0 0 0 0 0 316 316

1 1 1 3 2 7 28 30 57 67 72 96 111 49 48 573

0 0 0 0 0 0 0 0 0 0 1 0 1 0 2 0 2 1 7 1 10 2 15 4 32 1 22 0 481 564 573 573

Total

4th

1st

2nd

3rd 4th

0 0 0 0 0 0 0 0 0 0 0 1 0 2 570 573

1 1 1 4 4 16 54 43 71 102 110 153 196 67 66 889

0 0 0 0 0 1 1 3 3 11 15 21 54 33 747 889

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 3 0 4 1 1 0 1 2 878 886 889 889

Source: Authors.

V.2. We started supplying new products or services. V.3. We introduced new production or supply methods, such as CAD/ CAM, cell manufacturing systems, Internet marketing, or shortened distribution channels. V.4. We obtained new customers. V.5. We found new suppliers. V.6. We established new sections in charge of R&D or venture businesses. Question V.1 is related to industrial upgrading and Questions V.2–V.6 to innovation. We asked the above questions with regard to four different time periods: (1) Period I (January 2005–Sepetember 2007); (2) Period II (January 2002–December 2004); (3) Period III (January 1999–December 2001); and (4) Period IV (before 1998). Table 4.13 and Figure 4.1 indicate trends in upgrading and innovation, and Table 4.13 shows those numbers inside and outside a cluster and percentages divided by the total numbers of SME. Overall, the number of upgrades and innovation has been increasing, except with regard to upgrading in Period IV. Obtaining new customers (V.4), supply of new products and services (V.2), and introduction of new production or supply methods (V.3) show large increases in recent periods, and the SMEs analysed here have been attempting these activities intensively, with more than two-thirds of SMEs

126

Shoichi Miyahara and Masatsugu Tsuji

successful in achieving upgrading and innovations in Period I. In what follows, we analyse the background to these improvements in performance. On comparison of SMEs inside and outside a cluster, from Table 4.13 it follows that there are noticeable differences in these activities. Table 4.13 Number of upgrading and innovation: Replies to question V Period I (Jan. 2005 to Sept. 2007) Inside of cluster Outside of cluster

Upgraded business activities Supply of new products or services Introduction of new production or supply methods Obtaining new customers We found new suppliers Establishment of new sections in charge of R&D or venture businesses Did nothing above No reply Total

Freq.

%

72

0.23

187

Freq.

Total

%

Freq.

%

99

0.17

171

0.19

0.59

322

0.56

509

0.57

116

0.37

223

0.39

339

0.38

199 110 67

0.63 0.35 0.21

342 197 106

0.60 0.34 0.18

541 307 173

0.61 0.35 0.19

9 18 778

0.03 0.06 2.46

17 38 1344

0.03 0.07 2.35

26 56 2122

0.03 0.06 2.39

Period II (Jan. 2002 to Dec. 2004) Inside of cluster Outside of cluster

Upgraded business activities Supply of new products or services Introduction of new production or supply methods Obtaining new customers We found new suppliers Establishment of new sections in charge of R&D or venture businesses Did nothing above No reply Total

Total

Freq.

%

Freq.

%

Freq.

%

46 138

0.15 0.44

74 236

0.13 0.41

120 374

0.13 0.42

86

0.27

157

0.27

243

0.27

155 89 47

0.49 0.28 0.15

290 148 67

0.51 0.26 0.12

445 237 114

0.50 0.27 0.13

19 33 613

0.06 0.10 1.94

43 65 1080

0.08 0.11 1.88

62 98 1693

0.07 0.11 1.90

Continued

Relationship between Innovation of SMEs and Clustering

127

Table 4.13 Continued Period III (Jan. 1999 to Dec. 2001) Inside of cluster

Upgraded business activities Supply of new products or services Introduction of new production or supply methods Obtaining new customers We found new suppliers Establishment of new sections in charge of R&D or venture businesses Did nothing above No reply Total

Outside of cluster

Total

Freq.

%

Freq.

%

Freq.

%

29 97

0.09 0.31

68 176

0.12 0.31

97 273

0.11 0.31

50

0.16

102

0.18

152

0.17

137 65 33

0.43 0.21 0.10

231 120 44

0.40 0.21 0.08

368 185 77

0.41 0.21 0.09

49 55 515

0.16 0.17 1.63

83 100 924

0.14 0.17 1.61

132 155 1439

0.15 0.17 1.62

Period IV (before 1998) Inside of cluster Outside of cluster

Upgraded business activities Supply of new products or services Introduction of new production or supply methods Obtaining new customers We found new suppliers Establishment of new sections in charge of R&D or venture businesses Did nothing above No reply Total

Total

Freq.

%

Freq.

%

Freq.

%

38 80

0.12 0.25

70 151

0.12 0.26

108 231

0.12 0.26

38

0.12

81

0.14

119

0.13

105 63 28

0.33 0.20 0.09

188 96 35

0.33 0.17 0.06

293 159 63

0.33 0.18 0.07

64 67 483

0.20 0.21 1.53

111 133 865

0.19 0.23 1.51

175 200 1348

0.20 0.22 1.52

Source: Authors.

4.2.3.2 Innovation SMEs were also asked in question V.3 about their achievements related to three types of innovations, which are the number of patents applied for, those registered, and new products and services developed. The replies to those questions are summarized in Tables 4.14, 4.15 and 4.16 respectively.

128

Shoichi Miyahara and Masatsugu Tsuji

600 500 400 300 200 100 0 Period I (Jan. 2005 to Sept. 2007)

Period II (Jan. 2002 to Dec. 2004)

Upgraded business activities. Supply of new products or services. Introduction of new production or supply methods. Obtaining new customers. We found new suppliers. Establishment of new sections in charge of R&D or venture businesses. 400 350 300 250 200 150 100 50 0 Period III (Jan. 1999 to Dec. 2001)

Period IV (before 1998)

We upgraded business activities. We started supplying new products or services. We introduced new production or supply methods. We obtained new customers. We found new suppliers. We established new sections in charge of R&D or venture businesses. Figure 4.1

Trend of Upgrading and Innovation

Source: Authors.

Relationship between Innovation of SMEs and Clustering

Table 4.14 Number of patents applied for

0 Under 5 5–10 10–15 15–20 Over 20 No reply Total

Inside of cluster

Outside of cluster

Total

Freq.

%

Freq.

%

Freq.

%

105 85 18 4 1 0 103 316

33.23 26.90 5.70 1.27 0.32 0.00 32.59 100.00

194 141 21 6 4 2 205 573

33.86 24.61 3.66 1.05 0.70 0.35 35.78 100.00

299 226 39 10 5 2 308 889

33.63 25.42 4.39 1.12 0.56 0.22 34.65 100.00

Source: Authors.

Table 4.15 Number of patents registered Inside of cluster

0 Under 5 5–10 10–15 15–20 Over 20 No reply Total

Outside of cluster

Total

Freq.

%

Freq.

%

Freq.

%

133 61 2 1 1 1 117 316

42.09 19.30 0.63 0.32 0.32 0.32 37.03 100.00

231 102 10 4 0 1 225 573

40.31 17.80 1.75 0.70 0.00 0.17 39.27 100.00

364 163 12 5 1 2 342 889

40.94 18.34 1.35 0.56 0.11 0.22 38.47 100.00

Source: Authors.

Table 4.16 Number of new products and services developed Inside of cluster

0 Under 10 10–30 30–50 50–100 Over 100 No reply Total Source: Authors.

Outside of cluster

Freq.

%

Freq.

%

27 137 22 7 17 10 96 316

8.54 43.35 6.96 2.22 5.38 3.16 30.38 100.00

60 219 35 5 44 6 204 573

10.47 38.22 6.11 0.87 7.68 1.05 35.60 100.00

Total Freq. 87 356 57 12 61 16 300 889

% 9.79 40.04 6.41 1.35 6.86 1.80 33.75 100.00

129

130 Shoichi Miyahara and Masatsugu Tsuji

Tables 4.14 and 4.15 show that a higher percentage of SMEs inside a cluster achieved patents applied for and patents registered in the most recent three years than SMEs outside a cluster, especially in the class with under five times. Moreover, this characteristic is revealed much more clearly in regard to for the number of new products and services developed, which is indicated in Table 4.16. A higher percentage of SMEs inside a cluster experienced these products innovations than SMEs outside a cluster. 4.2.4 Summary of survey results From the above discussions, we can summarize the results of the mail survey in comparison of the two groups of SMEs inside and outside a cluster as follows: 1. Upgrading and innovation: In recent years, such as Periods I and II, relatively more SMEs inside a cluster experienced upgrading than those outside a cluster. On the other hand, in Periods III and IV, the latter achieved more than the former. This implies that recently SMEs inside a cluster became more active than before. Regarding various types of innovations shown in Table 4.13, there is no distinct difference, on the other hand, in terms of innovations defined by patents, the differences between two groups can be found. SMEs inside a cluster achieved more patents applied for, patents registered and new products and services developed than those outside a cluster in the most recent three years. 2. Characteristics of SMEs: Here we summarize the results of the mail survey. Originally, we selected 5,000 SMEs, including 2,000 inside and 3,000 outside a cluster to receive the questionnaire, and received 889 replies, 316 from inside and 573 from outside a cluster. The ratio of replies from the two samples is closely similar to that of the original sample. The characteristics of the two respondent groups of SMEs are also closely similar in firm size and category of industry. There are a number of differences in the variety of firms within the manufacturing sector. Moreover, other characteristics such as type of relationship with other firms, namely, self-products, keiretsu, or non-keiretsu, as well as business indicators, such as the amount of sales and profits, are closely similar among the two groups. 3. R&D investment: According to Table 4.11, two-thirds of SMEs in the two groups spent less than 5 per cent, including 0 per cent, of sales on R&D investment, which is the basis of upgrading and innovation. Even though R&D investment was small, they seem to have achieved reasonably good results in upgrading and innovation, as indicated by Table 4.11. This can be analysed by rigorous statistical methods in what follows. 4. R&D ratio and business performances: Two interesting observations were found in the results of the mail survey, namely that of the relationship between R&D ratio and trend of sales, and business performance. Table 4.17 indicates these relationships. More than 60 per cent of SMEs

Table 4.17 Ratio of R&D and sales trend Funds for R&D 0%

Decreasing Trend of sales amount Same within 3 years Increasing No reply Total

0–5%

5–19.9%

Over 20%

No reply

Total

Freq.

%

freq.

%

Freq.

%

Freq.

%

Freq.

%

Freq.

%

18 35 34

20.69 40.23 39.08 0.00

67 131 237 1

15.37 30.05 54.36 0.23

28 32 80

20.00 22.86 57.14 0.00

4 11 9

16.67 45.83 37.50 0.00

33 67 100 2

16.34 33.17 49.50 0.99

150 276 460 3

16.87 31.05 51.74 0.34

87

100.00

436

100.00

140

100.00

24

100.00

202

100.00

889

100.00

Source: Authors.

Table 4.18 Ratio of R&D and business performance Funds for R&D 0%

Balance of revenues and costs within 3 years

Source: Authors.

Surplus Balanced Deficit No reply Total

0–5%

5–19.9%

Over 20%

No reply

Total

Freq.

%

freq.

%

Freq.

%

Freq.

%

Freq.

%

Freq.

%

42 29 15 1 87

48.28 33.33 17.24 1.15 100.00

230 154 48 4 436

52.75 35.32 11.01 0.92 100.00

54 52 34

38.57 37.14 24.29 0.00 100.00

3 7 14

12.50 29.17 58.33 0.00 100.00

105 62 33 2 202

51.98 30.69 16.34 0.99 100.00

434 304 144 7 889

48.82 34.20 16.20 0.79 100.00

140

24

132 Shoichi Miyahara and Masatsugu Tsuji

with an R&D ratio of 0 per cent showed a decrease in or the same sales, and 17 per cent had negative profits. These numbers decrease as R&D ratio increases. However, the percentage of SMEs with an R&D ratio over 20 per cent in which sales were decreasing or the same was more than those with an R&D ratio of 0 per cent.6 The larger amount of R&D investment made their business worse, and a higher R&D ratio did not necessarily lead to greater sales. The same trend is applicable to business performance, which is shown in Table 4.18. These findings indicate that there is an optimal ratio of R&D.

4.3

Estimation of industrial clustering and innovation

Here we use rigorous econometric analysis to investigate the hypothesis that industrial clustering in certain regions promotes innovation. 4.3.1 Variables In the questionnaire, Question V asks SMEs about the results of industrial upgrading and innovations in different periods: Period I (January 2005–Sepetember 2007); Period II (January 2002–December 2004); Period III (January 1999–December 2001); and Period IV (before 1998). We choose dependent and independent variables for analysis from the questionnaire. 1. Dependant variables: Industrial upgrading in this chapter is defined according to several practices, namely from subcontracting simple works to producing intermediate goods, from producing intermediate goods to producing final products, or from simple to precise works. If SMEs experienced these kinds of upgrading during the above period, they reply ‘yes’. These replies of SMEs are taken as independent variables. The number of upgrading in four periods has already been shown in Table 4.13 and Figure 4.1. In addition, the questionnaire asks about innovations more concretely in the following way: (1) the number of patents applied for; (2) the number registered; and (3) the number of new products and services developed in the above four periods separately. The variables related to (1) and (2) were assigned a value of 0 if the SME did not experience innovation, 1 if the number was 1 to 4, and 2 if more than 5. The variable related to (3) was assigned a value of 0 if there no new products and services were developed, 1 if there were 1 to 9, and 2 if there were more than 10. These three kinds of numbers for each type of innovation are taken as dependent variables. Since three categories as defined earlier have rather small numbers in comparison with upgrading, which are shown in Tables 4.14, 4.15 and 4.16, we did not divide the period into four. The dependent variables in the innovation model are total numbers for whole time period.

Relationship between Innovation of SMEs and Clustering

133

2. Independent variables: The independent variables were as follows: (1) characteristics such as firm size in terms of the number of employees and the amount of capital, and the year of establishment; (2) business indicators such as sales amount and profits; (3) managerial orientation, which presents the attitude or behaviour of top management towards upgrading and innovation; (4) attitude of employees or organization towards upgrading and innovation; (5) location of SMEs inside or outside a cluster; and (6) distance from partners such as universities, regional research facilities, and other collaborating institutions. These are examples; the complete list of dependent variables is shown in the summary statistics in Table 4.19. 4.3.2 Models for estimations Since this chapter is focused on the relationships between industrial clusters and upgrading and innovation, emphasis is placed on mechanisms of the endogenous innovation process. In other words, how innovation or industrial upgrading are triggered by factors existing not only inside SMEs but also outside a cluster. The number of dependent variables is too great to construct models which include all dependent variables at one time, and accordingly we constructed sub-models for which dependent variables were selected according to the particular hypothesis under examination. 1. Upgrading model: The first hypothesis tested is the role of industrial clustering on SME upgrading, which is referred to as the ‘upgrading model’, by taking the number of upgrades in the four periods as dependent variable, which are represented by the following examples: (1) from being subcontractors for simple work to producing intermediate goods; (2) from producing intermediate goods to final products; and (3) from simple to complex or precision work. These are related to replies of Question V in the questionnaire. As for independent variables, we take the following SMEs’ characteristics: (1) year of establishment; (2) amount of capital; (3) number of employees; and (4) trend in sales amount in the most recent three years, and in addition, we consider the location of the SME inside or outside a cluster and distance from regional research facilities such as universities as independent variables. These last two variables are our main objectives. We adopt the Logit model for this analysis. 2. Innovation model: The second hypothesis tested is the relationship between innovation and industrial clustering. As dependent variables, we took data related to the number of (1) patents applied for, (2) patents registered, and (3) new products and services developed in the most recent three years, as enquired about in Question V.3. We refer to this as the ‘Innovation model’. We adopt the same variables as shown in the previous upgrading model. In this model, we use the Ordered Logit model for analysis, since these data are enquired about in three categories.

Table 4.19 Summary of statistics Obs.

Variable

Mean

Std. Dev.

Min

Max

Dependent variable V

1

1

V

1

V

1

Upgraded:

from Jan. 2005 to Sept. 2007

845

0.208

2

from Jan. 2002 to Dec. 2004

802

3

from Jan. 1999 to Dec. 2001

743

before 1988

0.406

0

1

0.151

0.358

0

1

0.133

0.34

0

1

134

V

1

4

698

0.16

0.367

0

1

V

3

2

Number of patents applied

902

1.116

0.849

0

2

V

3

3

902

1.054

0.911

0

2

V

3

4

Number of patents registered Number of new products and services developed

902

1.417

0.651

0

2

I

1

Year of establishment

876

1659

2007

Independent variable I 2 Amount of capital (logarithm) I

3

I

3

I

3

I

3

I

3

I

7

Number of employees (part-time employees working at least 8 hours are counted as one full-time employee)

890

7.472

34.345 0.968

4.382027 11.72126

from 4 to 9

902

0.175

0.38

0

1

from 10 to 19

902

0.216

0.412

0

1

from 20 to 49

902

0.283

0.451

0

1

from 50 to 99

902

0.166

0.373

0

1

100 and over

902

0.078

0.268

0

1

902

0.519

0.5

0

1

(Takes 1 if locate inside the cluster, takes 0 otherwise)

902

0.365

0.482

0

1

Sales trend over the past three years Inside or outside of industrial cluster

1969

135

IV

1

1

IV

1

2

IV

1

3

IV

1

4

IV

1

5

IV

1

6

IV

1

7

IV

1

8

IV

1

9

IV

1

10

The top management of your company:

pays attention to how well employees work together. demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals. keeps employees informed about management/ company policies and developments. encourages employees to expand their skill sets. promotes competition among employees. accumulates data on past successes and failures.

876

4.059

0.764

1

5

873

3.425

0.888

1

5

875

3.679

0.88

1

5

874

3.709

0.848

1

5

878

4.018

0.736

1

5

874

4.021

0.726

1

5

873

3.969

0.842

1

5

872

3.54

0.78

1

5

875

3.187

0.85

1

5

875

3.465

0.904

1

5

Continued

Table 4.19 Continued Variable 1

11

IV

1

12

IV

2

1

IV

2

2

IV

2

3

IV

2

4

IV

2

5

IV

2

6

136

IV

The management:

encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. challenges itself with new ideas and methods. places more emphasis on creating new technologies than updating existing ones. introduces new products faster than competitors. invests most of the budget in R&D. puts more effort into selling existing products than doing R&D for new ones. considers changes in the business environment opportunities rather than threats.

Obs.

Mean

Std. Dev.

Min

Max

871

3.56

0.873

1

5

873

3.803

0.911

1

5

883

4.101

0.753

1

5

876

3.411

0.854

1

5

875

3.707

0.967

1

5

871

2.921

1.104

1

5

873

2.951

0.864

1

5

871

3.821

0.814

1

5

137

IV

2

7

IV

2

8

IV

3

1

IV

3

2

IV

3

3

IV

3

4

IV

3

5

IV

3

6

IV

3

7

IV

3

8

IV

3

9

Your employees or organization:

adopts new strategies faster than competitors. makes decisions by looking forward and anticipating future business environments. considers employees’ spontaneous learning to be an important factor in company development makes efforts to analyse the successes and failures of past projects. always analyses competitors. attempts to study not only core technology but also other related types. are able to act on their own, without orders from the management. is discussed extensively among employees. is discussed extensively within management. understand what they should do. understand the company’s direction.

875

3.689

0.828

1

5

875

3.889

0.734

2

5

875

3.999

0.84

1

5

872

3.495

0.87

1

5

871

3.046

0.897

1

5

871

3.443

0.835

1

5

873

3.479

0.853

1

5

872

3.288

0.856

1

5

872

3.399

0.838

1

5

874

3.618

0.781

1

5

875

3.655

0.787

1

5

Continued

Table 4.19 Continued Variable

138

IV

3

10

VII

1

VII

6

1

VII

6

2

VII

6

3

recognize that the development of new business is important for the future of the company. Is the area where your company is located an industrial cluster, that is, are other companies, business groups, or universities also located there? How many minutes does it take within 30 minutes by car (or equivalent) to reach from 30 minutes to partners (other companies, 1 hour universities, and research from 1 to 1.5 hours institutes) with whom you from 1.5 to 2 hours collaborate on new projects?

Obs.

Mean

Std. Dev.

Min

Max

871

3.61

0.877

1

5

902

0.325

0.469

0

1

656

0.337

0.473

0

1

656

0.345

0.476

0

1

656

0.066

0.248

0

1

VII

6

4

656

0.095

0.293

0

1

VII

6

5

over 2 hours

656

0.122

0.327

0

1

VII

6

6

too far to go by car

656

0.037

0.188

0

1

Source: Authors.

Relationship between Innovation of SMEs and Clustering

4.4

139

Results of estimation I: Upgrading model

We present the results for the estimations of upgrading and innovation separately, beginning with upgrading. 4.4.1

Cluster and research facilities as factors in upgrading

In this estimation, we make an attempt to identify exact factors which promote industrial upgrading of SME during the different time periods, and we take characteristics of SMEs such as year of establishment, size of SMEs in terms of the amount of capital and the number of employees, sales trends, and attitudes of top management towards business operations as independent variables, and especially we focus on factors such as whether they are located inside or outside a cluster and distances from collaborating partners (other companies, universities and research institutes). The two main objectives of this chapter are to examine the role of clustering and collaborations with regional research institutions. We estimate the following models by choosing suitable independent variables, since there are many factors we asked in the questionnaire: (1) Case (a); (2) Case (b); and (3) Case (c). Case (a) includes variables except those related to distances from collaborating partners, Case (b) is referred to as the full model which contains all variables, while Case (c) is as the selected model in which variables are selected from the full model according to Akaike Information Criteria (AIC). The detailed estimated coefficients and statistics are shown in the Appendix. For making discussion clear, the estimation results of the upgrading model are summarized in Table 4.20, in which only results of Case (c) is presented and signs and significance levels of estimated coefficients are described. In what follows, we discuss our fundamental hypotheses such as the relationship of industrial clustering and collaborations with regional research institutions with industrial upgrading. 1. Industrial cluster: Let us examine the effects of industrial cluster and distance from regional research facilities such as universities. Location inside an industrial cluster is significant only in Period I at a 5 per cent level, which implies that in the recent five years clustering has become effective in promoting the upgrading of SMEs. This is an interesting and positive result to our hypothesis. This effect of upgrading is referred to the ‘clustering effect’. 2. Research facilities: The result of distance from research facilities and upgrading obtained by our estimation indicates quite interesting and important observations. At first, in Period IV (before 1998), ‘30–60 minutes distance (by automobile)’ is found to be related with upgrading significantly with the 10 per cent level, and its value of coefficient ‘30–60 minutes distance’, as shown in Table A1–4 of the Appendix, is over 0.412 for Case (c), which is quite large and actually the second largest in comparison with those of other significant variables. The distance of ‘60–90 minutes’ is negatively significant with upgrading in Period III (1999–2001) at the 10 per cent level, and weakly

Table 4.20 Summary of estimation results: Upgrading model Period I (Jan. Period II (Jan. Period III 2005 to Sept. 2002 to Dec. (Jan. 1999 to Period IV 2007) 2004) Dec. 2001) (before 1998) Year of establishment Amount of capital (logarithm) Number of employees: from 4 to 9 from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top management of pays attention to how well your company: employees work together. demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals. keeps employees informed about management/company policies and developments.

+ *

[**]

**

** + *

[**] [**] [+]

[*] + **

**

**

+

[*]

[**]

+

+ *

[**] +

Distance from collaborating partners (other companies, universities, and research institutes) Constant

encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour from 1 to 1.5 hours from 1.5 to 2 hours

[*] [**]

[+]

[+]

**

**

**

** [**] *

[+]

[*] +

over 2 hours too far to go by car [**]

[**]

Notes: [ ] indicates that its coefficient is negative. ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively. Source: Authors.

142

Shoichi Miyahara and Masatsugu Tsuji

20 per cent significant in Period II (2002–2004), and those coefficients are −0.895 and −1.705, respectively, and those are also large. Since the longer distance from these facilities provides a weaker relationship, these negative signs are reasonable. In addition, since we take the different distances as not continuous, but dummy variables, we can identify the exact distance within which these partners can collaborate with each other. From the fact that the signs of ‘1–1.5 hours’ by automobile are significantly negative in Periods II and III, it follows that research facilities located within this distance are rather too far to affect the upgrading of SMEs. In other words, distances of within 1 hour seems to be the critical point to upgrading. 4.4.2 Other factors We next examine the results of other factors which promote the industrial upgrading of SMEs. 1. Firm size: Table 4.20 identifies other factors which influence SME upgrading. Let us summarize the results of the upgrading estimation in what follows. First, industrial upgrading is related to year of establishment; SMEs established in Period I have weakly positive relationships with upgrading, while those established in Period II are negatively significant within the 5 per cent level. Secondly, the amount of capital is positive at the 10 per cent significance level in Period I, which implies that larger SMEs have experienced more upgrading in this period. Thirdly, a staff count of 10–19 employees is identified as a positive factor in Periods I and III, and these SMEs tend to have experienced more upgrading than those of other sizes. In Period IV, on the other hand, smaller SMEs in terms of employees have negative signs at the 5 per cent level. Table 4.20 also shows that the largest category of SMEs have a negative sign at the 10 per cent significance level. From these discussions, it is difficult to derive a general characteristic between upgrading and firm size, but at the early stage of agglomeration smaller SMEs tend to upgrade more than larger ones, while at the later stage the larger ones tend to experience upgrading.7 2. Managerial orientations: Variables related to managerial orientations of top of SMEs are categorized into three groups: those affected significantly to earlier stages of agglomeration, those affected to later stages, and those for the whole period. Variables such as ‘checks quality of work severely (negative)’, and ‘takes the leadership role in the planning of new business (negative)’ affected significantly to the early period (Period IV), while those such as ‘demands that employees follow routine procedures’, ‘listens to employees’ ideas and proposals’ and ‘promotes competition among employees (negative)’ affected to recent Periods I and II. Top management’s behaviour such as ‘interested in employees’ experience for nurturing’, ‘encourages employees to take risks and challenge themselves’ and ‘accumulates data on past successes and failures (negative)’ are significantly related to upgrading for all periods.8

Relationship between Innovation of SMEs and Clustering

143

It is also difficult to derive a certain positive conclusion from these findings, but among them, ‘encourages employees to take risks and challenge themselves’ is significant for all periods at a 5 per cent level, which is found to be most important for managerial orientation for upgrading. 4.4.3 Summary of the upgrading model In this section, we interpret the above estimation results in asmuch as how the upgrading processes have evolved in accordance with agglomeration. At the early stage of agglomeration (Period IV), smaller SMEs experienced upgrading and this upgrading might be promoted by the collaboration with regional research institutions which were closely located to SMEs. At this stage, there is no clear difference among SMEs located inside or outside a cluster. At the current period (Period I) in which the certain level of agglomeration is achieved, SMEs which have the following characteristics are experiencing upgrading; those with large capital with larger growth of recent sales, and located in the industrial clusters. Regional research intuitions are found to be in no significant relationship with their upgrading. In other words, we can conclude that the clustering effect is larger than the collaboration effect, or the upgrading model might describe the process alternatively in such a way that the reasons of upgrading switch from collaborations to clustering.

4.5

Results of estimation II: Innovation model

Here we present the results of analysis using the innovation model by focusing on hypotheses on factors promoting innovation among SMEs, especially effects of clustering and regional collaborations towards innovation. 4.5.1 Innovation and research facilities as factors in innovation This chapter focuses on innovation of the following three types: (1) Case (a) patents applied for; (2) Case (b) patents registered; and (3) Case (c) new products and services developed. SMEs were asked to provide these numbers for the most recent three years, which are dependent variables in the innovation model, while independent variables are exactly the same as the previous upgrading model. We do not repeat details such as summary statistics of independent variables here again. We analysed three models according to the above types of innovation using the Ordered Logit model, since the number of replies to each of the three questions was categorized into three classes.9 The results of these analyses are summarized in Table 4.21. Let us examine our hypotheses in what follows. 1. Industrial cluster: Table 4.21 indicates that location inside an industrial cluster is negatively significant to innovation at the 20 per cent level in Case (a) (patents applied) and Case (c) (new products and services developed), but

Table 4.21 Summary of estimation results: Innovation model Number of patents applied for Year of establishment Amount of capital (logarithm) Number of employees:

from 4 to 9 from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top management of your pays attention to how well employees company: work together. demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals. keeps employees informed about management/company policies and developments.

** **

Number of Number of new patents products and registered services developed ** + [**]

[**] [**]

[**] [+]

[**] [+]

[+]

+

**

[**] [**]

[+]

[+] ** [*]

**

**

Distance from collaborating partners (other companies, universities, and research institutes)

encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour from 1 to 1.5 hours from 1.5 to 2 hours over 2 hours too far to go by car

[+]

**

+ +

[*] [+]

[**] [*]

Notes: [ ] indicates that its coefficient is negative. ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively. Source: Authors.

146

Shoichi Miyahara and Masatsugu Tsuji

is not significant in Case (b) (patents registered). These findings indicate that SMEs inside a cluster tend to achieve less innovation, but because of low significance levels, the relationships are weak. 2. Regional research institutions: In Table 4.21, significant variables related to the number of three categories of innovations are ‘over 2 hours’ for Case (b) with the 10 per cent significance level and Case (c) with the 5 per cent level, and ‘too far to go by car’ for Case (a) with the 20 per cent level and for Case (c) with 10 per cent level. In all cases, these coefficients are negative. They thus show that the longer the distance between the SMEs and regional research facilities, the weaker the relationship becomes. The results obtained in the innovation model coincide with realty. In Case (c), a 2-hour distance is the maximum according to the results. In sum, the results of this analysis reveal that clustering has negatively related to innovations by SMEs, and SMEs inside a cluster are less innovative. Regarding the relationships with regional research institutes, SMEs located further from these institutions are less innovative. The collaboration effect is larger than the clustering effect in the innovation model. 4.5.2 Other factors 1. Firm size: A clear relationship between firm size and innovation can be found from the results. In Cases (a) and (b), larger firms with from 20–99 employees have negative signs at the 20 per cent level, which indicates that among SMEs with 20–99 employees, smaller ones tend to experienced more innovation related to patents applied and registered than larger ones. Innovations seem to be carried out by smaller companies. In Case (c), in addition to SMEs with 20–99, the smallest class of SMEs is negatively significant at the 5 per cent level, and the same assertions as Case (a) and (b) are applicable. As for firm size in terms of capital, Case (a) has a positive sign at the 5 per cent level, while Case B is at the 20 per cent level, which implies that in both cases larger SMEs tend to carry out innovations. This is reasonable, since innovations require a large amount capital and larger SMEs can afford to finance them. 2. Managerial orientations: In this case, fewer variables are significant in comparison with the upgrading model. A common factor to all three cases is ‘listens to employees’ ideas and proposals’, which is positively significant at the 5 per cent level. It is reasonable that innovation requires creativity of employees. Other common replies to Case (a) and (b) are ‘demands that employees follow routine procedures’ and ‘encourages employees to take risks and challenge themselves’, and these two are positively significant. These seem to present strong management. The variables with negative signs, on the other hand, are ‘keeps employees informed about management/company policies and developments’ and ‘promote competition among employees’ for Case (a), and ‘gives power and responsibility to the offices’ for Case (b).

Relationship between Innovation of SMEs and Clustering

147

It is rather difficult to explain why these have negative relationships with innovation, but they are related to leadership.10 In Case (c), no other variable influences the development of new products and services. 3. Year of establishment: Table 4.21 shows that in Cases (a) and (b), the year of establishment has a positive coefficient, which implies that SMEs with a longer history tend to achieve more innovations; in other words, start-ups achieve fewer innovations than existing SMEs. It should be noted, however, that here we are discussing the number of patents only, without reference to patent quality.11 4.5.3 Summary of the innovation model The estimation of results within the innovation model is not as complicated as the upgrading model, since the number of significant variables are small and the pattern of their signs do not contradict each other. Let us summarize the innovation model case by case. 1. Case (a) patents applied: Innovations in this category are carried by SMEs with rather long histories and with larger sizes in terms of capital. For SMEs with the number of employees of 20–99, innovation is conducted by the rather smaller-sized SMEs. Regarding the recent trend in sales, SMEs with declining sales trends tend to have more patents applied, which can be interpreted in that they put more effort into innovation for surviving against the competition. SMEs inside a cluster tend to have fewer patents applied than those outside a cluster, which seems to be contradict reality; in other words, the model fails to verify the hypothesis. As for managerial orientations, SMEs with ‘demands that employees follow routine procedures’, ‘listens to employees’ ideas and proposals’ and ‘encourages employees to take risks and challenge themselves’ have more patents applied for, while SMEs that replied yes to ‘keeps employees informed about management/company policies and developments’ and ‘promotes competition among employees’ tend to have fewer patents applied for. Finally, SMEs located too far to go by car have fewer patents applied, which coincides with reality. 2. Case (b) patents registered: For this category of innovation, years of establishment, firm size in terms of capital and employees, and sales trend are close to Case (a). Location inside or outside a cluster is not significant. As for managerial orientations, the same items with positive relationships as Case (a) are also significant to patents registered, while ‘give power and responsibility to the offices’ is negatively significant to patents registered. Finally, SMEs located over 2 hours away by car have fewer patents registered, which coincides with reality. 3. Case (c) new products and services developed: Innovation in this category does not have significant relationships with SMEs’ year of establishment and firm size with respect to capital. Firm sizes such as from 4 to 9 and from 20 to 99 are negatively significant; that is, SMEs of a smaller size tend have more new

148

Shoichi Miyahara and Masatsugu Tsuji

products and services. SMEs inside a cluster tend to have fewer new products and services developed than those outside a cluster, which seems to contradict reality. As for managerial orientations, SMEs with ‘listens to employees’ ideas and proposals’ and ‘takes the leadership role in the planning of new business’ conduct more new products and services developed. Finally, SMEs located over 2 hours away by car and those located too far to go by car have fewer new products and services developed, which again coincides with reality.

4.6 Conclusions Based on an extensive mail survey, this chapter provides a number of new insights into upgrading and innovation achieved by Japanese SMEs. The results confirm the role of industrial clustering as a factor in promoting upgrading and innovation, particularly with regard to the most recent five years. SMEs inside a cluster achieved better performances in terms of frequencies of upgrading and numbers of innovations in comparison with those outside a cluster. This is consistent with reality, in that the national and regional governments have made great efforts in areas such as deregulation and funding. The margin of differences in all accounts is not great, rather small. The results of econometric analysis show the differences with a small significance level. There remain problems yet to be solved.12 Several issues related to these models warrant further research. First, we do not include the characteristics of each cluster. Regions have different regional resources which are incorporated in the upgrading and innovation of regional SMEs; that is, the number of universities, junior colleges, banks, legal offices, college students, industrial structures, and so on.13 Secondly, policy measures are important to the fostering of regional clusters, and this leads to the question of strategy to encourage the development of industrial clusters. In the global context, Kuchiki and Tsuji (2005, 2008), Tsuji et al. (2008), Tsuji et al. (2006), and Tsuji and Ueki (2008) have proposed and verified the so-called ‘Flowchart Approach’ to agglomeration which successfully explains the recent growth of East Asian industrial clusters. One of the final objectives of the present chapter is to apply analysis to the development of consistent policies for building successful industrial clusters. Thirdly, the process how accumulated tacit information and knowledge among agents inside a cluster turned to be implicit upgrading and innovation, and this process is not expressed in terms of any mathematical equations or any solid formula.14 In addition to those mentioned, there are several methodological ways to expand the analysis presented in this chapter. First, there are many questions which we did not handle, such as ‘new products and services’ and ‘new production and marketing methods’. These items are thought to be outcome of innovation. In order to apply the Logit or Ordered Probit model, we categorize the number of patents or upgrading, for example, in such a way that a variable takes 0 if there is none, 1 if there are 1–4, or 2 if over 5. The numbers themselves can be taken as variables.

Table A1–1 Results of estimation: Upgrading model I (1) Period I (Jan. 2005 to Sept. 2007) Case (a) Coefficient

149

Year of establishment Amount of capital (logarithm) Number of from 4 to 9 employees: from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top pays attention to how well management of employees work together. your company: demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals.

Case (b)

t-value + +

Coefficient

t-value

0.005 0.184 −0.016

1.470 1.720 −0.040

0.446 −0.136 0.089 −0.758 0.289 0.382 0.044

1.100 −0.330 0.200 −1.320 1.510 2.010 0.300

0.286

2.310

−0.084

−0.650

0.005 0.169 −0.013

1.500 1.590 −0.030

0.462 −0.112 0.103 −0.708 0.283 0.384 0.060

1.150 −0.270 0.230 −1.240 1.490 2.040 0.410

0.277

2.270

−0.076

−0.590

0.326

2.330

**

0.304

−0.297

−1.850

*

0.150

0.900

+ **

**

Case (c) Coefficient t-value + *

0.005 0.178

1.550 1.790

+ *

0.473

2.190

**

+ + **

−0.746 0.263 0.372

−1.760 1.410 1.990

* + **

**

0.267

2.410

**

2.170

**

0.362

2.670

**

−0.263

−1.630

+

−0.268

−1.720

*

0.148

0.880

0.207

1.300

+

Continued

Table A1–1 Continued Case (a)

150

keeps employees informed about management/company policies and developments. encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour from 1 to 1.5 hours from 1.5 to 2 hours

Distance from collaborating partners (other companies, universities, and over 2 hours too far to go by car research institutes) Constant Nob Log likelihood Rseudo R2

Case (b)

Coefficient

t-value

0.087

0.610

0.092

0.640

0.053

0.380

0.036

0.260

−0.232

−1.860

*

−0.228

−1.820

*

−0.220

−1.830

*

−0.283

−2.300

**

−0.283

−2.280

**

−0.294

−2.480

**

0.256

1.900

*

0.254

1.870

*

0.303

2.380

**

0.107

0.920

0.109

0.930

−0.316 −0.443 −0.097

−1.350 −0.940 −0.260

−0.044 −0.859

−0.140 −1.310

+

−14.075 770 −369.379

−2.160

**

−14.028 773 −373.418

−2.230

**

−14.239 770 −371.410 0.059

−2.200

Coefficient

Case (c)

**

0.064

Note: ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively.

t-value

Coefficient t-value

+

0.056

Table A1–2 Results of estimation: Upgrading model II (2) Period II (Jan. 2002 to Dec. 2004) Case (a)

151

Year of establishment Amount of capital (logarithm) Number of from 4 to 9 employees: from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top pays attention to how management of well employees work your company: together. demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices.

Case (b)

Coefficient

t-value

Coefficient

t-value

−0.001 0.061 0.624

−0.340 0.480 1.130

−0.001 0.078 0.623

−0.500 0.610 1.130

0.837 0.610 0.475 −0.025 −0.145 0.043 −0.053

1.570 1.140 0.830 −0.040 −0.670 0.190 −0.320

0.813 0.591 0.441 −0.060 −0.113 0.000 −0.063

1.520 1.100 0.770 −0.090 −0.520 0.000 −0.380

0.152

1.100

0.136

0.970

−0.027

−0.180

−0.036

−0.240

0.131

0.810

0.129

0.800

−0.571

−3.120

−0.548

−2.960

+

**

Case (c) Coefficient t-value −0.002

−4.150

**

−0.481

−2.790

**

+

**

Continued

Table A1–2 Continued Case (a) Coefficient

152 Distance from collaborating partners (other companies, universities, and research institutes)

listens to employees’ ideas and proposals. keeps employees informed about management/ company policies and developments. encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour from 1 to 1.5 hours from 1.5 to 2 hours

Case (b)

t-value

Coefficient

Case (c)

t-value

Coefficient t-value

0.304

1.600

+

0.299

1.560

+

0.296

1.620

+

0.233

1.440

+

0.235

1.440

+

0.282

1.860

*

0.177

1.090

0.163

1.000

−0.070

−0.480

−0.061

−0.410

−0.237

−1.670

*

−0.222

−1.550

+

−0.213

−1.650

+

0.522

3.190

**

0.512

3.110

**

0.508

3.420

**

−0.102

−0.770

−0.090

−0.670

−0.141

−0.550

−0.992 −0.574

−1.570 −1.140

+

−0.895

−1.450

+

over 2 hours too far to go by car

153

Constant Nob Log likelihood Rseudo R2

−2.447 731 −298.348 0.048

−0.430

−0.305 −0.642 −1.438 731 −295.890 0.060

Note: ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively.

−0.770 −0.820 −0.250 742 −303.320

Table A1–3 Results of estimation: Upgrading model III (3) Period III (Jan. 1999 to Dec. 2001) Case (a) Coefficient

154

Year of establishment Amount of capital (logarithm) Number of from 4 to 9 employees: from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top pays attention to how well management of employees work together. your company: demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals.

Case (b)

t-value

0.005 −0.066 0.358

1.210 −0.480 0.500

1.147 0.945 1.235 0.859 −0.059 −0.261 −0.043

1.750 1.430 1.790 1.070 −0.250 −1.030 −0.230

−0.076

Coefficient

t-value

0.005 −0.034 0.370

1.090 −0.240 0.520

1.165 0.938 1.239 0.853 −0.051 −0.283 −0.041

1.760 * 1.410 + 1.790 * 1.060 −0.210 −1.110 −0.210

−0.510

−0.097

−0.640

0.094

0.580

0.091

0.560

0.247

1.360

0.225

1.230

* + *

+

−0.003

−0.020

0.038

0.180

0.042

0.190

0.050

0.230

Case (c) Coefficient t-value

0.771 0.442 0.635

2.390 1.390 1.840

** + *

0.249

1.620

+

155

keeps employees informed about management/ company policies and developments. encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour from 1 to 1.5 hours

Distance from collaborating partners (other from 1.5 to 2 hours companies, universities, and research institutes) over 2 hours too far to go by car Constant Nob Log likelihood Rseudo R2

0.210

1.140

0.192

1.030

−0.016

−0.090

−0.029

−0.170

−0.183

−1.130

−0.179

−1.100

−0.198

−1.250

−0.187

−1.180

0.363

2.060

**

0.372

2.100

−0.183

−1.280

+

−0.182

−1.270

−0.281 −1.791

−0.960 −1.730

−0.350

−0.680

−13.661 679 −251.104 0.048

−1.580

+

0.107 −0.367 −12.859 679 −247.995 0.060

Note: ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively.

−0.182

−1.310

+

**

0.336

2.160

**

*

−1.705

−1.660

*

−3.797 710 −267.887 0.041

−5.330

**

0.270 −0.470 −1.480 +

Table A1–4

Results of estimation: Upgrading model IV

(4) Period IV (before 1998) Case (a)

156

Year of establishment Amount of capital (logarithm) Number of from 4 to 9 employees: from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top pays attention to how well management of employees work together. your company: demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals.

Coefficient

t-value

0.000 −0.167 −1.126

0.000 −1.200 −2.050

−0.641 −0.406 −0.060 −0.415 −0.067 0.010 0.142

−1.330 −0.870 −0.120 −0.650 −0.290 0.040 0.790

−0.138

−0.940

−0.287

−1.890

0.198

Case (b) Coefficient

Case (c)

t-value

Coefficient t-value

−0.001 −0.181 −1.128

−0.160 −1.270 −2.040 **

−0.638 −0.398 −0.038 −0.393 −0.043 0.014 0.116

−1.310 −0.850 −0.080 −0.610 −0.190 0.060 0.630

−0.137

−0.920

−0.282

−1.850

1.120

0.189

1.060

0.137

0.650

0.114

0.540

−0.297

−1.430

−0.277

−1.320

** +

*

+

−0.218 −0.914

−2.940 −2.310

** **

+

−0.364

−1.310

+

*

−0.307

−2.470

**

0.237

1.610

+

+

157

keeps employees informed about management/company policies and developments. encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. Distance from from 30 minutes to 1 hour collaborating from 1 to 1.5 hours partners (other from 1.5 to 2 hours companies, universities, and research institutes) over 2 hours too far to go by car Constant Nob Log likelihood Rseudo R2

0.058

0.330

0.061

0.350

0.080

0.470

0.116

0.680

0.151

0.940

0.139

0.860

0.069

0.440

0.047

0.290

0.321

1.800

*

0.345

1.910

*

0.352

2.270

**

−0.291

−2.100

**

−0.292

−2.110

**

−0.283

−2.220

**

0.442 −0.582 0.542

1.660 −0.900 1.300

*

0.412

1.700

*

+

0.579

1.450

+

0.272 −0.180 0.542 635 −262.391 0.062

0.700 −0.230 0.080

−0.398 635 −265.233 0.052

−0.060

Note: ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively.

657 −273.976

Table A2–1 Results of estimation: Innovation model I (1) Number of patents applied Case (a) Coefficient

158

Year of establishment Amount of capital (logarithm) Number of from 4 to 9 employees: from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top pays attention to how well management of employees work together. your company: demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals.

Case (b)

t-value

0.005 0.167 0.009

2.510 2.140 0.030

0.156 −0.232 −0.242 0.318 −0.261 −0.165 −0.026

0.570 −0.840 −0.790 0.850 −1.930 −1.210 −0.250

0.100

Coefficient ** **

t-value

0.006 0.167 −0.009

2.520 2.130 −0.030

0.136 −0.260 −0.277 0.283 −0.250 −0.176 −0.021

0.490 −0.940 −0.910 0.760 −1.850 −1.280 −0.210

1.150

0.099

1.140

−0.026

−0.290

−0.037

−0.400

−0.047

−0.490

−0.046

−0.470

−0.073

−0.640

−0.071

−0.620

0.286

2.470

0.283

2.440

*

**

Case (c) Coefficient ** **

*

**

t-value

0.005 0.203

2.210 2.890

** **

−0.392 −0.413

−2.520 −2.180

** **

−0.260 −0.183

−1.970 −1.340

** +

0.107

1.410

+

0.259

2.510

**

159

keeps employees informed about management/ company policies and developments. encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour

Distance from collaborating from 1 to 1.5 hours partners (other from 1.5 to 2 hours companies, universities, and research institutes) over 2 hours too far to go by car /cut1 /cut2 Nob Log likelihood Rseudo R2

−0.160

−1.610

0.047

0.480

−0.123

−1.350

0.088

1.010

0.192

1.970

0.043

0.520

12.126 13.266 820 −861.644 0.026

+

+

**

−0.150

−1.510

0.035

0.360

−0.117

−1.280

0.091

1.030

0.179

1.820

0.048

0.580

−0.027

−0.170

0.071

0.230

−0.253

−0.960

−0.212 −0.503 12.077 13.220 820 −860.091 0.028

−0.880 −1.270

Note: ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively.

+

*

−0.170

−1.880

*

−0.110

−1.290

+

0.213

2.510

**

−0.504 10.341 11.479 825 −866.752 0.026

−1.300

+

Table A2–2 Results of estimation: Innovation model II (2) Number of patents registered Case (a) Coefficient Year of establishment Amount of capital (logarithm)

t-value

0.004 0.078

1.980 1.010

from 4 to 9

−0.053

from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top pays attention to how well management of employees work together. your company: demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals.

Number of employees:

Case (b) Coefficient **

t-value

160

0.005 0.080

2.060 1.030

−0.180

−0.068

−0.240

0.125 −0.235 −0.192 0.174 −0.170 −0.153 0.029

0.440 −0.830 −0.620 0.450 −1.250 −1.100 0.280

0.088 −0.252 −0.223 0.155 −0.162 −0.172 0.029

0.310 −0.890 −0.720 0.400 −1.190 −1.230 0.280

0.155

1.780

0.155

1.770

−0.040

−0.440

−0.052

−0.560

−0.062

−0.630

−0.053

−0.530

−0.111

−0.960

−0.118

−1.010

0.263

2.250

0.261

2.220

*

**

Case (c) Coefficient **

*

**

t-value

0.005 0.099

2.190 1.430

** +

−0.320 −0.280

−2.050 −1.460

** +

−0.176

−1.310

+

0.152

2.030

**

−0.145

−1.320

+

0.265

2.420

**

Continued

161

keeps employees informed about management/ company policies and developments. encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour

Distance from collaborating from 1 to 1.5 hours partners (other from 1.5 to 2 hours companies, universities, and research institutes) over 2 hours too far to go by car /cut1 /cut2 Nob Log likelihood Rseudo R2

−0.046

−0.460

−0.033

−0.320

0.018

0.180

0.016

0.160

−0.083

−0.910

−0.073

−0.790

0.113

1.270

0.121

1.350

0.110

1.110

0.093

0.940

0.074

0.880

0.078

0.920

−0.068

−0.410

0.023

0.080

−0.208

−0.780

−0.467 −0.217 10.581 11.316 820 −828.769 0.023

−1.940 −0.540

10.249 10.981 820 −830.934 0.020

Note: ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively.

+

*

0.135

1.620

+

−0.424

−1.850

*

10.863 11.591 824 −834.930 0.020

Table A2–3 Results of estimation: Innovation model III (3) Number of new products and services developed Case (a)

162

Year of establishment Amount of capital (logarithm) Number of from 4 to 9 employees: from 10 to 19 from 20 to 49 from 50 to 99 100 and over Sales trend over the past three years Inside or outside of industrial cluster The top pays attention to how well management of employees work together. your company: demands that employees follow routine procedures. checks quality of work severely. is interested in employees’ experience for nurturing. gives power and responsibility to the offices. listens to employees’ ideas and proposals.

Coefficient

t-value

0.000 0.028 −0.552

−0.090 0.350 −1.850

−0.275 −0.545 −0.576 −0.130 −0.147 −0.172 0.049

−0.930 −1.860 −1.780 −0.330 −1.050 −1.210 0.470

0.051

Case (b)

Case (c)

Coefficient t-value

Coefficient

0.000 0.029 −0.567

0.040 0.370 −1.900

−0.312 −0.580 −0.613 −0.181 −0.138 −0.174 0.050

−1.050 −1.960 −1.880 −0.460 −0.980 −1.220 0.470

0.570

0.058

0.650

0.029

0.310

0.015

0.150

−0.135

−1.340

−0.140

−1.390

+

0.146

1.240

0.155

1.310

+

0.219

1.820

0.211

1.740

*

*

* *

+

*

t-value

*

−0.403

−2.120

**

* *

−0.349 −0.424

−2.090 ** −2.160 **

−0.193

−1.390

+

0.284

2.970

**

163

keeps employees informed about management/ company policies and developments. encourages employees to expand their skill set. promotes competition among employees. accumulates data on past successes and failures. encourages employees to take risks and challenge themselves. takes the leadership role in the planning of new business. from 30 minutes to 1 hour

Distance from collaborating from 1 to 1.5 hours partners (other from 1.5 to 2 hours companies, universities, and research institutes) over 2 hours too far to go by car /cut1 /cut2 Nob Log likelihood Rseudo R2

0.020

0.200

0.043

0.410

−0.100

−0.960

−0.115

−1.110

0.032

0.350

0.038

0.410

−0.055

−0.610

−0.051

−0.560

0.080

0.800

0.070

0.700

0.099

1.150

0.104

1.200

−0.143

−0.850

−1.312 1.059 820 −749.029 0.018

0.102

0.310

−0.085

−0.310

−0.462 −0.867 −0.897 1.491 820 −745.174 0.023

−1.880 −2.080

Note: ***, **, * and + stand for the significance level at the 1%, 5%, 10% and 20%, respectively.

0.112

* **

−0.448 −0.755 −1.166 1.216 860 −783.451 0.019

1.450

+

−1.970 ** −1.860 *

164

Shoichi Miyahara and Masatsugu Tsuji

Acknowledgement In the preparing the mail survey, the authors are indebted to the supports of the Small and Medium Enterprise Agency for obtaining information on SMEs authorized by the New Business Promotion Act for SMEs. The financial support provided by Japan Society for the Promotion of Science is gratefully acknowledged.

Notes 1. Regarding endogenous innovation process, see Tsuji (2005) and Tsuji et al. (2008), for example. 2. The same mail surveys for enquiring about industrial upgrading and innovations in East Asia were conducted in October–November 2007 in Indonesia, the Philippines, Thailand and Vietnam, and the results are found in Tsuji and Ueki (2008). 3. The New Business Promotion Act for SMEs, which was legislated in 1999. A total of 30,931 firms have been authorized as of December 2007. 4. Areas referred to as a ‘cluster’ accord with those defined by the Regional Industry Activating Act, which was legislated in 1997. The aim of this Act is to strengthen the basis of regional economies by promoting clustering industries in the region. 5. SMEs in Ohta ward in metropolitan Tokyo tend to supply their products to Keiretsu; in contrast, those in Higashi-Osaka are more commonly independent manufacturers. See Tsuji et al. (2005) and Bunno et al. (2006). 6. With regard to this point, no clear difference is found between SMEs inside and outside a cluster. 7. This can be interpreted as follows: at the early stage of agglomeration, rather smaller SMEs started upgrading, while after the certain level of agglomeration larger SMEs started experiencing upgrading. In other words, In accordance with agglomeration, firm size has also been expanding, and this induces industrial upgrading. In this sense, at the early stages of agglomeration, small SMEs experience upgrading. This promotion of upgrading due to expanding firm size induces upgrading is referred to as the ‘scale effect’. Thus two causes of upgrading, that is, the ‘size effect’ and ‘clustering effect’. Form the discussion of the size effect, it follows that in accordance with agglomeration, upgrading occurred from small to large SMEs. It is, therefore, reasonable that the agglomeration promotes increase in firm size as well as upgrading. We thus have to separate this process of upgrading and the clustering effect. This is one point for future research. 8. Tsuji et al. (2005) also found that managerial orientation of ‘Accumulates data on past successes and failures’ has negative effect to ICT use by SMEs. This can be interpreted that top managements with the strong leadership of starting some new business activities established rather their own way of decision-making. 9. The construction of these dummy variables is explained in 4.3.1 in Section 4.3. 10. We obtained the result that this variable has a positive influence on the introduction of IT (Information Technology) to SMEs in Bunno et al. (2006). 11. It is difficult to measure the quality of innovations. In the mail survey, we simply enquired about the number, without mentioning quality. 12. It is said that 70 per cent of authorized SMEs are not satisfied with the current New Business Promotion Act for SMEs.

Relationship between Innovation of SMEs and Clustering

165

13. Imagawa (2005, 2007), for instance, uses these regional resources, including the number of restaurants, to estimate factors related to clustering in the IT industry. 14. This process can be expressed in the form of stochastic differential equations. See Fujita and Thisse (2002).

References Bunno, T., M. Tsuji, H. Idota, H. Miyoshi, M. Ogawa and M. Nakanishi (2006) ‘An Empirical Analysis of Indices and Factors of ICT Use by Small and Medium-sized Enterprises in Japan.’ Proceedings of ITS Biennial Conference (CD-ROM), Beijing, China. Fujita, M., P. Krugman, and A. Venables (1999) The Special Economy: Cities, Region, and International Trade. Cambridge, MA: MIT Press. Fujita, M. and J.-F. Thisse (2002) Economics Agglomeration: Cities. Industrial Location, and Regional Growth. Cambridge: Cambridge University Press. Imagawa, T. (2005) ‘Japan: Remedies to Activate Local Cities,’ in Kuchiki and Tsuji (2005), pp. 299–318. —— (2007) ‘Information Technology and Economic Growth: Discovering the Informational Role of Density,’ in Tsuji, Giovannetti, and Kagami (2007). Krugman, P. (1991) Geography and Trades, Cambridge, MA: MIT Press. Kuchiki (2007) ‘Agglomeration of Exporting Firms in Industrial Zones in Northern Vietnam: Players and Institutions,’ in Tsuji, Giovannetti, and Kagami (2007), pp. 97–138. Kuchiki, A. and M. Tsuji (eds) (2005) Industrial Clusters in Asia: Analyses of their Competition and Cooperation. Basingstoke: Palgrave Macmillan. —— (2008) The Flowchart Approach to Industrial Cluster Policy. Basingstoke: Palgrave Macmillan. Porter, M. E. (1980) The Competitive Advantage of Nations. New York: Free Press. Saxenian, A. L. (1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Cambridge, MA: Harvard University Press. Tsuji, M. (2005) ‘Country Report of Japan,’ in Information Technology for Development of Small and Medium-sized Exporters in Latin America and East Asia, M. Kuwayama, Y. Ueki and M. Tsuji Eds, pp. 345–374, UNDP/ECLAC/IDE-JETRO, October. Tsuji, M., E. Giovannetti, and M. Kagami (eds) (2007) Industrial Agglomeration and New Technologies: A Global Perspective. Cheltenham: Edward Elgar. Tsuji, M., S. Miyajara, and Y. Ueki (2008) ‘An Empirical Examination of the Flowchart Approach to Industrial Clustering: Case Study of Greater Bangkok, Thailand,’ in Kuchiki and Tsuji (eds) (2008), pp. 194–261. Tsuji, M., S. Miyahara, Y. Ueki, and K. Somrote (2006) ‘An Empirical Examination of Factors Promoting Industrial Clustering in Greater Bangkok, Thailand,’ Proceedings of 10th International Convention of the East Asian Economic Association (CD-ROM), Beijing, China. Tsuji, M., H. Miyoshi, T. Bunno, H. Idota, M. Ogawa, M. Nakanishi, E. Tsutsumi, and N. Smith (2005) ‘ICT Use by SMEs in Japan: A Comparative Study of Higashi-Osaka and Ohta Ward, Tokyo.’ OSIPP Discussion Paper, No. 06–05, Osaka University. Tsuji, M. and Y. Ueki (2008) ‘Consolidated Multi-countries Analysis of Agglomeration,’ in Analyses of Industrial Agglomeration, Production Networks and FDI Promotion, M. Ariff Ed., ERIA Research Project Report 2007 No. 3, ERIA (Economic Research Institute for ASEAN and East Asia).

5 Collective Goods for Reformatting the Rio de Janeiro Software Cluster into a Local Innovation System Antonio José Junqueira Botelho, Alex da Silva Alves and Glaudson Mosqueira Bastos

5.1

Introduction

The flowchart model has significantly contributed to the policy knowledge about the development of manufacturing industrial agglomerations in emerging economies in Asia into high-productivity, efficient clusters (Kuchiki 2004; Kuchiki and Tsuji 2005b). The flowchart model recent research efforts have been targeted at firm geographical agglomerations in other regions such as North and Latin America, particularly Brazil and the United States, and other types of industries such as renewable – sugar-alcohol production chain (Ueki 2007) – and non-renewable energy – oil and gas exploration and production (Botelho and Bastos, this volume) and software in India (Okada 2005; Okada, this volume); as well as exploring new topics such as innovation clusters in the United States and China (Kabir et al. 2007; Kuchiki 2007). This chapter represents yet another unique addition to this second phase of the flowchart research programme, as it studies the conditions for the redevelopment of a non-hierarchical cluster in the IT sector centred on innovation, with a particular focus on software and services industry (SSI), in the metropolitan region of Rio de Janeiro, Brazil. Estimates of the size of the Rio de Janeiro IT cluster vary widely from 500 to 1,250 to 5,000. The number of non-micro one- to two-person firms in the more circumscribed software and services industry (SSI) sector amounts to 1,250, of which about 500 software houses, which together represent 16 per cent of the state’s IT market. Most firms are small (fewer than ten employees), constituting a decentralized, non-hierarchical cluster, although in a few segments there is a significant number of firms of more than 30 employees. The state SSI industry association ASSESPRO RJ estimates that the top 20 firms have on average over 500 employees, and account for 15,000 to 18,000 employees of the total employment of 50,000. The balance of about 166

Rio de Janeiro Software Cluster 167

30,000 employees is accounted for by 17,000 one- to two-employee micro enterprises. Although there are no hard numbers available there is a firm consensus among local software industrial associations and development agencies that Rio de Janeiro software and services industry (SSI) has been in a relative decline since its heyday in the 1980s. ASSESPRO RJ estimates that the state’s share of the national industry was halved in a decade, going from 36 per cent in 1997 to 18 per cent in 2006, but it remains the country’s second largest. The share of the total number of IT firms which represented 18 per cent (5,022) in 1991 remains stagnated by 2001, although in absolute numbers it follows the total growth rate of 280 per cent over the period (ASSESPRO RJ 2005). Diverse empirical elements appear to sustain this uneasy feeling. One such indication is that Rio de Janeiro state software registry in the National Intellectual Property Institute (INPI) relative ranking has been declining over the past two decades. Furthermore, whereas São Paulo registries almost doubled between 1988–1999 and 2000–2006 and those of the third ranked state in the latter period (Minas Gerais) tripled, Rio de Janeiro registry ranking has fallen to number two after dividing the leadership with São Paulo and the absolute number of registrations has grown less than 30 per cent between the periods. Next, in the past couple of years a dozen SSI firms have made IPOs to finance their expansion to acquire a scale to compete become internationally, none founded or headquartered in Rio de Janeiro. Finally, as subsidiaries of large SSI firms in Brazil have stepped up their offshore outsourcing activities, most of this expansion has taken place in São Paulo – One of IBM’s and Brazil’s largest software development centre with 4,000 employees has been located in Hortolandia (in IBM’s old manufacturing facilities), nearby São Paulo capital,1 and EDS opened in late 2007 its largest single US$20 million software development services centre with 5,000 of its 10,000 employees in the country2 in the city of São Bernardo do Campo, in the greater São Paulo metropolitan region. The promotion of the territorial agglomeration of high-tech firms in local production systems of SMEs can be sustained by external economies and the collective goods that induce them. In particular, external economies that favour the generation of new products instead of their reproduction (in scale) to the marketplace. In high-tech sectors, the production process – once an output is finished – is relatively simple and much less costly in terms of time, work employed and space required than it tends to be in the case of industrial districts. Once produced something that ‘works’, meaning there is demand for it in the marketplace, its production can be started in an industrial scale at relatively lower costs. The major problem then lies in the inputs structure, that is to say, in the generation capacity of new products in sectors where this process tends to be strongly influenced by scientific advances. That argument is the main motivation behind the explanation of the importance of external economies linked to formal and

168

Antonio José Junqueira Botelho et al.

informal relationships with scientific institutions as well as with the ‘production’ of specialized services that favour the connection between, on one side, education and scientific structures and, on the other, entrepreneurs and other stakeholders. Paraphrasing Bagnasco’s argument on the socioeconomic success of north and north-east Italian industrial districts in the 1970s, it could now be said that there is also a process of social construction of innovation, given the growing importance of the local sphere for innovation activities. In sum, external economies favouring the link between the generation process of new knowledge and the launching of new products seem to be much more important for small and medium-sized firms in high-tech sectors. In the generation of collective goods necessary for the evolution of this one particular form of production organization, the intentional processes that give birth to the promotion of cooperation among collective public and private agents are more important than those collective goods rooted in the original endowments of a territory. More importantly, it is the nature of social capital that differs among industrial and technological districts. Social capital is in the root of the generation process of collective goods. Collective goods produce the external economies that enhance the conditions for the agglomeration of firms in a territory. What thus explains the difference between those external economies arising out as the result of intentional processes (e.g., geared at the solution of a given problem) to those built through a process of long-standing trust founded on local identities is social capital. In technological districts, social capital tends to be formed by experimentation rather than by the construction of a local identity – or embeddedness. The evidence on the long-standing success behind the Silicon Valley seems to be oriented towards a cyclical dynamics of start-up creation and spin-offs from universities, public and private research labs and other firms that – even though most of them do not last too long – promote a sort of sane contagion of the local innovation culture. Behind this phenomenon there is a professional community – without walls or physical borders – working overnight towards the solution of business and research problems arising out by high-tech entrepreneurs’ decisions to risk in new ventures (Saxenian 1994). Additionally, even before the establishment of such communities, as Porter (2001) and Picchieri (2002) suggest, a group of intermediate organizations formed by the interaction of public agents in the local, regional and national spheres needs to be strengthened. The intermediate organizations constitute an important collective good needed to favour the link between research and the business environment (Antonelli 2001; Cooke 2001; Crouch et al. 2004; Granovetter et al. 2000; Kenney 2000; Saxenian 1994 and Storper 1997). They collectively suggest that the origins of technological districts seem to be given, on the one hand, by incremental and

Rio de Janeiro Software Cluster 169

spontaneous processes developed in terms of previous local competencies and resources (collective goods); on the other, and parallel to the prior aspect, there are also strong evidences indicating that without a qualifying public support – local and national – the development of technological districts would have become almost impossible. Innovation clusters tend to be more dependent on conscious political choices geared at establishing support organizations which stimulate cooperation between, on one side, educational and research institutions and, on the other side, firms and society at large. The production of these collective goods induces important external economies. These external economies, conversely, favour the localization of new small and medium enterprises and promote new positive externalities linked to the agglomeration effects that take place in both traditional and technological districts. Thus, public policy does represent an important role in the evolution of technological districts, not by securing a more efficient resource allocation in the local economy by means of subsided loans, tax exemptions, fiscal incentives, etc. such as public policies designed for traditional industrial districts development. The role of public support in technological districts tends to be much more promotional for the dynamics of such environments, being targeted to promoting the building up of intermediate organizations, such as science parks, business incubators, technology transfer offices as well as the (indirect) support to the formation of services organizations and associations as diverse as clubs of angel investors. This chapter focuses on a particular type of local production system, characterized by a complex set of SMEs operating in high-technology sectors such as, to name a few, software, biotechnology and media. Among the many classifications and names given to this phenomenon, these systems tend to be termed, particularly in Europe, as technological districts or innovation clusters. The chapter is divided as follows: The next section discusses the analytic imbrications of topics of local development, innovation and clusters. The following section briefly presents the Brazilian SSI industry and characterizes in greater detail Rio de Janeiro SSI evolution, discussing the basis of its strengths and weaknesses and identifying barriers and opportunities for future growth. The next section discusses the role of innovation in developing high-tech clusters, suggesting that critical modifications are needed in the flowchart model to apply it to the SSI in the case of Brazil, which are to substitute the anchor firm for local governance in the cluster phase and the lead scientist for collective learning culture in the innovation phase of the model (see Figure 5.1). The fourth section presents the results of an empirical survey with a small sample of SSI firms aimed at testing the proposition of the importance of collective learning and other conditions convergent with the flowchart model for reformatting the SSI cluster in Rio de Janeiro into an innovation framework capable of putting it into a new sustainable growth path.

170 Antonio José Junqueira Botelho et al.

Step 1 Agglomeration

Industrial zone (software market segments: finance, telecom, retail and oil & gas) Capacity-building Local governance (Rio Software Network, SEPRORJ and ASSESPRO)

Decentralized, non-hierarchical agglomeration (proto)cluster

Related firms

Step 2 Innovation

Universities and research institutes

Capacity-building

Lead scientist for collective learning culture Innovation-based cluster Figure 5.1

A flowchart of IT cluster in the City of Rio de Janeiro

Source: Authors’ own development based on Kuchiki and Tsuji (2005a).

The concluding remarks raise some policy considerations and further details the suggested modifications in the flowchart model to account for SSI in emerging economies.

5.2 Local economic development, clusters and innovation3 New technologies following an uncertain and not yet consolidated trajectory are paving the way for the formation of high-tech systems fundamentally constituted by small and medium-sized enterprises (SMEs) (Keeble and Wilkinson 1999; Storper 1997; Swann et al. 1998). Research efforts have been directed towards identifying the variables explaining the motivations behind the agglomeration of firms in a territory, where scientific knowledge constitutes an important input in its production organization and, given that innovation is acknowledged as the most desired output of such territorial systems, local economic development strategies have been deemed as important to contribute to spur innovation among SMEs in a local production system (OECD 1996). It is assumed that an SME-based dynamism is more secure for local communities exploring certain production specializations because, as the literature suggests, it tends to be embedded in the ability of

Rio de Janeiro Software Cluster 171

territorial actors to sustain and to reproduce specialized knowledge in the production of certain products and services. Local production systems and, in particular, SMEs within these systems are the ones to benefit the most in a territorial context when competitiveness, quality job creation and social cohesion are ensured by the presence of solid collective goods. 5.2.1 Policy-relevant characteristics of local economic development Local economic development regards the capacity of local institutional agents to cooperate in order to start and to conduct a regional development agenda capable of mobilizing local and external resources and competences. The protagonists of local economic development efforts are usually those agents who can coordinate a set of vertical and horizontal initiatives capable of attracting external resources of political (qualified public investments or resources for bringing the private sector into the territory), as well as of cultural and economic nature (bound to investment decisions or to the localization of private agents). It is important to distinguish local economic development from local dynamism. The latter is measured merely in terms of income and employment generated in a territory as the consequence of a given policy (or political actions). Local economic development efforts, on the other hand, are about using external resources to bring value to the local assets of a territory in order to attract investments, external firms, cultural and scientific structures not only as a means to promote increases in production, income and employment, but fundamentally to enrich local competences and specializations (Antonelli 2001; Becattini 2000; Sapir et al. 2003). Thus, local economic development policies aim at qualifying a given social and economic environment with targeted interventions so that to increase the availability of infrastructure and services, as well as to foster cooperation of firms within their production processes (Trigilia 2005). There is an extraordinary variety of production systems’ linkages (Storper 1997). Local production systems can be very different from one another, not only concerning their productive apparatus, but also with reference to the social structures of which they are constituted. The events by which each community has built its own set of values are very diverse and dynamic so that there are countless axes around which communities, even of recent formation, find cohesion and solidarity.4 As a main building block of regional policies, local production systems permeate the whole socioeconomic fabric of a community, thus putting together governments, firms, local (and national) institutions and individuals in a coordinated effort aiming at the promotion of quality of life, infrastructure building, job creation and the competitiveness of firms operating in a territory. 5.2.2 External economies The external economies approach stresses the role of increasing returns within circumscribed regional spaces to which firms have access because

172

Antonio José Junqueira Botelho et al.

of the important role of proximity. Consequently, externalities stem from imperfect divisibilities among production factors so that proximity provides enhanced opportunities for agents to internalize their benefits (Antonelli 1986; Brusco 1992; Camagni 1999; Piore and Sabel 1984). External economies can be the result of material (cooperation) and even of non-material collective goods, as can be the case of external economies made possible by the implementation of Information and Communication Technologies. External economies can be considered as the fruit of local collective goods which contribute to increasing the competitiveness of firms operating in a given territory. They can reduce costs and improve technological innovation in firms, particularly SMEs. The reason is that firms cannot produce on their own – or are not even interested in doing so given potential free riding effects – in a significant amount the quantity and quality of such goods that might be needed for improving their competitiveness conditions (Crouch et al. 2004). In certain cases such goods can be of a fully public nature, such as the availability of qualified workforce, good communications and logistics infrastructure; and, in other cases, they can be of exclusive access to certain groups as it can be the case of recycling infrastructures available to certain non-environmentally friendly production sectors. Nevertheless, external economies cannot take place without the production (and reproduction) of solid collective goods. In the case of technological districts, collective goods have a different and complementary nature in respect to industrial districts, in part explained by the type of external economies that need to be induced for the socioeconomic take-up of such environments and in part by the dynamics of technological development and innovation. The understanding of such differences require a broad analysis of the local character of innovation and the role played by the territory for the development of small and medium firms operating in high-technology segments. 5.2.3 Making innovation local A local innovation system (LIS) is built upon local social structures and institutions, thereby more carefully reflecting the development of knowledge and competencies within a regional or local community. For Bagnasco and Sabel (1995, p. 55), in local production systems: ‘1. the relationship among businesses are characterized by a close interweaving of competition and cooperation; 2. the relationships between entrepreneurs and their employees – both at micro-level within the business and at the macro-level within industrial relations – present at any one time elements of conflict and elements of participation; 3. the productive and, more generally, the social structure are rich in knowledge closely connected with production, technology, marketing and often financial administration and management.’ Along these lines, an innovation cluster thus refers to the specific

Rio de Janeiro Software Cluster 173

geographical localization and the cultural fabric of the innovation system in which it is embedded. An LIS focuses on the interests of a community or region, thus exploring local interests and competencies (Antonelli 2001), being more in line with local demands of the territory. All such localized environments, thence, are bound to specific aspects of the communities to which they belong, of which the most resilient one is social capital. A version of technological districts regards university-based local innovation systems. Here, the university is attributed a leading role, meaning that its R&D efforts and its technological partnerships with companies act as structuring and guiding elements in the innovative activity and, therefore, in opportunities to generate new enterprises. The new organizational arrangement, formed by new and emerging businesses, companies engaged in joint R&D with universities and research universities, constitutes the basis of a university-based LIS, together with market organizations (with their specialized equipment suppliers, services and customers) and non-market organizations (universities, research institutes, local trade associations, regulatory agencies, technology-transfer offices, business associations, relevant government agencies, etc.). So far, the results of public-led interventions for the development of LIS, though, tend to be in many accounts disappointing both in developed and in developing countries, although innovation (or the need to innovate) is today very present in the policy agenda of many nations. Most such policy to promote the technological capacity-building of firms and to build up solid national innovation systems, have been so far very much influenced by institutional and deterministic views of technological development (Leyten 2004; May 2002; Navarro 2003), neglecting other factors with a more ‘localized’ nature, such as the levels of quality of life, security and urban infrastructure of a territory which strongly influence the attractiveness of qualified human capital (Amendola et al. 2005). 5.2.4 External economies and collective goods The mobility, openness and flexibility of markets brought about by globalization also introduces new opportunities so that effective local economic development strategies that incorporate such opportunities to increasing the economic value of local resources need to be strongly emphasized. The acquisition of technological capabilities by SMEs tends to be strictly related to the quality of the provision of collective goods to SMEs. It is usually assumed that high-tech is whatever incorporates new technologies strictly connected to scientific advances, which is measured in terms of high rates of R&D expenditures – including scientific personnel employed in these activities. Among the sectors more frequently included in this category, there can be mentioned the pharmaceutical and chemical industries, aerospace and related industries, the biotechnology sector and the IT industry. Furthermore, most high-tech industry is not organized in local

174

Antonio José Junqueira Botelho et al.

production systems of small and medium-sized enterprises, being each of them followed by differing dynamics, history and evolution processes. The high-tech sectors more commonly associated with local production systems of SMEs – as evidenced by empirical comparative research works – tend to be the ones in the biotechnologies, media production and information and communications technologies (in particular software development). The question that intrigues researchers the most, everywhere, lies on understanding the determinants of the decisions of firms in these industries to organize their production around decentralized systems (mainly) characterized by small and medium-sized enterprises. It seems though that the literature on industrial districts provides a good indication in this regard. As the Becattinian tradition suggests, a necessary prerequisite for the development of local production systems of SMEs lies in the divisibility of the production process. Another two aspects can be added here: (1) the uncertainty underlying technological paths, in this sense making more convenient a business approach founded on the experimentation among different players (external to the firm), as the experience with biotechnology suggests (DeVol and Bedroussian 2006; Powell 1996); (2) the constant market variations (in demand structure, in regulation, etc.) that requires a continuous flexible recombination of production factors, as experiences in the media business (movies, TV, etc.) and software development are putting in evidence (Bresnahan et al. 2001; Castells 2000). However, given that not all high-tech sectors give life to local production systems of SMEs, a further qualification seems necessary. In high-tech systems of SMEs there are also significant variations. As an example, not all software-houses (developers) and Internet companies can be seen as high-tech firms, given the low resilience of some of these companies on scientific knowledge and advances. The utilization of widespread methods, such as the number of employees engaged in R&D activities, tend to have reduced significance for supporting the classification of SMEs operating in these sectors. In addition to this difficulty, there are also high levels of innovations not captured by traditional indicators. This differentiation is important because it may directly affect the success of public policy and the types of collective goods available to firms in a technological district. The main explanation for the decisions of firms in high-tech sectors to organize in a territory rather than make use of the advantages brought about by globalization to reduce costs by recurring to outsourcing (as many firms in traditional industrial districts do), is in Becattini’s (1979) interpretation that takes into account the tangible and intangible external economies of which local firms take advantage in their utilization, together with the types of collective goods produced (intentionally or not). This approach, however, provides just part of the answer as it was originally developed to explain why firms in traditional districts do agglomerate. Thus, we hypothesize that the

Rio de Janeiro Software Cluster 175

external economies in innovation clusters tend to differ from those originated in traditional industrial districts. Understanding these differences is important due to the policy design mistakes and implementation error, as well as the consequent utilization by territorial actors (mainly firms) of the collective goods generated as the result of innovation and regional policies. Three main aspects need to be considered. First, it is important to consider the territorial actors’ conditions of access to the knowledge generated by research enterprise as well as the availability of communication (and interaction) channels with scientific and university facilities. This is actually a fundamental collective good that guarantees a continuum of technological upgrading and capacity-building or upgrading of firms as well as the availability of an evergreen flow of qualified personnel. The presence of each of these institutions may vary from case to case, although they are strictly interdependent and represent, together, a fundamental role in the development of high-tech local production systems. Regarding the first external economy (technological upgrading) made available by this collective good, the possibility of formal relationships (contracts, joint ventures) among companies (or groups of firms) and the research institutions must be ensured by the rule of law so to guarantee that contracts (mainly in terms of confidentiality agreements) will be met and the intellectual property will be secured. Here, scientific and educational institutions also influence the territory’s production system by means of more informal relations through personal networks that put firms together with the research centres’ environments. This type of external economy is more important than the first given the influence exerted by innovation activities on the competitiveness of firms. In this sense is usually formed a group of ‘professional communities’ that are particularly relevant for the exchange of information, for the development of modes of tacit knowledge and local trust, as well as for head-hunting of qualified professionals. The sort of social capital generated is established in terms of ‘experimentation’ and weak ties, rather than on a social construction process of trust and the collective sense of embeddedness, as it does in traditional industrial districts. Secondly, the availability of qualified suppliers of goods and services for firms constitutes another relevant collective good in these clusters. This refers to external economies formed as the result of emergent effects of (mostly) non-intentional processes. Such processes tend to be bound to the original localization of some firms as well as educational and research structures that successively induce the evolution of entrepreneurial and quality workforce resources. This element is of significant importance to the modes of production presupposing a specialized division of labour and a horizontal integration of small and medium-sized enterprises. It is not, of course, an established rule to all firms in the high-tech sector. In the software sector this high division of labour and flexible organization of production tends

176

Antonio José Junqueira Botelho et al.

to be very much pronounced (Botelho et al. 2005; Biagiotti and Burroni, in Crouch et al. 2004; Ramella and Trigilia 2006). In general, though, one cannot imagine such networks as autarchic closed systems which do not interact with the external environment. Networks not only interact but can be also largely be influenced by the external setting. Demand for high-tech products and services usually comes from the markets outside their territorial boundaries (Bresnahan et al. 2001). Brazilian software clusters, for example, tend to explore its domestic market, with very low penetration rates in international segments, mostly due to losing industrial policies in the past decades and macroeconomic uncertainties throughout the 1980s and 1990s, thereby affecting long-term market expansion strategies of firms (Botelho 2005; De Negri and Salerno 2005). Markets for most of the firms operating in an industrial district tend to be the firms operating in the final edge of the same district’s specialization value-chain. Finally, in LIS, there are usually significant cooperation ties with large firms, most of them external to the district. Nevertheless, as the empirical evidence suggests, the availability of local partners bound to formal and informal cooperation ties is presented as an important condition for the operation of firms, thus influencing (and qualifying) the overall system dynamics. Beyond services associated with the valorizations of research and education, a very important role is played by financial, marketing and entrepreneurs’ support services (coaching to start-ups). The technological path leading to the launching of innovative products and services takes time to mature and is uncertain and risky, so that the presence of these intermediary service providers is growing in importance, particularly regarding specialized financial services such as business angels and venture capitalists. The territorial embeddedness of such financial institutions, frequently taking place by means of transfers of experienced and qualified individuals from the research to the business environment, and from industry to the finance environment is also very important, because it paves the way for more coherent (and less risky) evaluations on the feasibility of technological projects (Kenney 2000; Powell et al. 2002). Without quality financial services that very strongly contribute to bridging the knowledge-to-business process, even good ideas founded on prospective market opportunities cannot transform into the assets that so importantly contribute to local economic development. Therefore, it can be assumed that proximity is important for the development of high-tech activities, because it favours the formation of tacit knowledge and its utilization through direct, face-to-face interactions among local agents in the process of generating innovations. In sectors where a technological path has not yet been consolidated, firms tend to agglomerate in a territory in order to exploit their synergies: the growth of tacit knowledge and the enhanced chances of participating in innovative networks are

Rio de Janeiro Software Cluster 177

important externalities that may help explain the territorial concentration of small and medium-sized high-tech firms. However, externalities related to the generation of new technical knowledge seem more important for high-tech local systems. In this respect, it can be hypothesized that the influence of such externalities – or untraded interdependencies in the words of Storper (1997) – is increased in those activities more strongly based on the continuous generation of new knowledge and on scientific advances. In addition, changes in the market and in consumers’ preferences brought about a need for greater flexibility in production processes. In so far as production is customized, such as in the media industry, and even customized over a broad scope of functions and continuously over time, as in certain software developments, the need for flexibility is higher, and this can explain the presence of many small firms that collaborate at a local level, taking advantage of external economies. Trust and social networks play an important role in reducing transaction costs for sharing valuable information in productive interactions exposed to opportunistic behaviour such as moral hazard and free riding. However, in LIS, it is then hypothesized that trust is less embedded in local identities historically built in a territory and more rooted within communities formed around the high-calibre competencies of individuals. They are developed by local agents in their careers through different firms and research and university institutions. Given the stronger role of a highly educated labour force that is usually much more volatile and less loyal to a firm or to a territorial identity, the capacity of the territory to attract – and to maintain – qualified individuals by providing high salaries, quality of life and very good infrastructure counts for much more than local history and other social and political features of a territory. On the cognitive side, a key role is played by research and university structures, in that they can provide the territory with a myriad of collective goods capable of enriching it with positive externalities. For in high-tech systems, universities and research institutions, together with the R&D facilities of large firms, provide qualified labour and chances of formal collaborations and informal exchange between specialized actors. The same applies to the role of financial institutions (like Venture Capitalists and business angels) and other specialized business services (like management consulting, business and technological coaching). In this regard, local and regional governments can influence the formation of high-tech systems with appropriate policies that enhance the chances of developing a localized base for technical knowledge production (and reproduction) so that the fine-tuning of an adequate channel (or infrastructure) to promote communication with research institutions and industry becomes fundamental. LIS constitute a different type of local production system, being fundamentally structured towards understanding the governance dynamics of these systems, given the confusion developed by policy-making wings in

178 Antonio José Junqueira Botelho et al.

terms of effective policy design and implementation to favour technological systems of SMEs, particularly in middle-income developing countries whose policy recipes follow those designed in industrialized countries. These clusters are sustained by specific collective goods which, in turn, are produced through the interplay with complementary governance institutions at local, regional and national levels. References made to the governance dimension can thus help analyse the relationship between national conditions and local factors, thus allowing the combination of useful insights coming from the national innovation systems approach with those offered by the studies of industrial districts and of localized technical knowledge. It is widely accepted that these collective goods are determinants for the success of LIS. Nonetheless, there seems to be some sort of disregard to a third type of collective good that produces significant (positive) external economies: the quality of the environment. The quality of the environment surrounding and within a local production system reflects the capacity of institutional stakeholders to (intentionally) produce collective goods by means of quality cooperation. Naturally, the availability of an adequate infrastructure at affordable costs is fundamental for SMEs, particularly those in their early stages and start-up growth phases. In this regard, there are varying types of technological and business incubators and science parks in such high-tech local production systems. Not of lesser importance is the availability of good communications infrastructure supporting easy links to national and international centres. The importance of these nodes, mainly founded on a quality infrastructure provision, induces the territory to value its internal resources thus attracting qualifying investments and other external stakeholders. Although these resources tend to be a prerequisite to successful industrial districts and other local production systems, it seems that there is an additional peculiarity to consider regarding technological districts. In the latter, it seems that socio-cultural and environmental quality also plays an important role, of lesser relevance in other local production systems. This particular factor influences the capacity of a territory to attract (and to maintain and to subsequently renew) qualified and specialized workers who come to a territory together with their families. This sole fact conditions the chances for the establishment of relatively stable ‘professional communities’ that are in the roots of the experimentation processes leading to LIS.

5.3

Rio de Janeiro IT cluster

The Brazilian SSI market has almost doubled in size between 2004 and 2007, growing 23 per cent in 2007. Services growth has been higher (24 per cent) than software (20 per cent). In 2006 the Brazilian SSI market reached US$9.1 billion (out of a total US$16 billion IT market, 1.3 per cent of the global market), of which US$5.8 billion in services and US$3.3 billion in software

Rio de Janeiro Software Cluster 179

(ABES 2007), the world’s thirteenth largest and representing almost 1 per cent of Brazil’s GDP. Domestic software production accounted for over US$1 billion (33 per cent of the market), of which 24 per cent was standard, 72 per cent was customized and less than 5 per cent was for export of licenses (US$52 million). Custom software exhibits the largest growth rate, 36 per cent, higher than the overall SSI (23 per cent) and services alone (24 per cent). Service exports were close to US$200 million. According to this source almost 8,000 firms (over two-thirds in software, 31 per cent in development and 69 per cent in distribution) were involved in the development, production and distribution of software and in service provision and of those in development and production, 94 per cent were micro and small.5 Overall, small firms are 57 per cent and micro 37 per cent of the total number of firms. The largest software market segments are industry (onequarter), finance (one-fifth) and services, including telecom and related activities (one-seventh) but the fastest-growing segments are agribusiness (95 per cent), retail (61 per cent) and oil and gas (55 per cent). Since 1995, according to the RJ software business association (SEPRORJ) the software market has been growing at an average annual rate of 11 per cent, three times larger than the hardware segment. This rate has been accelerating in recent years with the diffusion of IT and the shift in business models towards increased outsourcing. Thus the software and services market grew 15 per cent in 2005 and 15.4 per cent in 2006, and is expected to grow 14 per cent in 2007. The city of Rio de Janeiro has a population of 6 million people (2007) and the country’s second largest GDP of about US$66 billion (2005, 5.5 per cent of total and 3.3 per cent of population), less than half of that of São Paulo, the largest (12.3 per cent of total and 6 per cent of population), and 50 per cent larger than the third biggest, Brasilia, the country’s capital. The top five municipal (out of 5,564 municipalities) GDPs account for one-quarter of Brazil’s GDP (versus top ten in 1985). Services represent 66 per cent of Brazil’s GDP. Rio de Janeiro’s GDP fell to 5.5 per cent of Brazil’s; continuing evidence of the city’s historical economic decline started with the move of the nation’s capital to Brasília in 1960. Although Rio de Janeiro represents just 2.8 per cent of the industrial GDP versus almost 10 per cent for São Paulo, Rio de Janeiro and São Paulo together account for over 20 per cent of the country’s services GDP. Although all major capitals have experienced a decline in their share of the services GDP over the period 2002–2005, Rio de Janeiro’s decline at almost 1 per cent was the largest among them, reaching 6.5 per cent of this total in 2005 (7.3 per cent in 2002). The Rio de Janeiro metropolitan region (2007) has a population of 11 million and the state 15.4 million. 5.3.1 Historical evolution Rio de Janeiro was the birthplace of information technology in Brazil, as it had the first computer in the country to process census data back in 1960.

180

Antonio José Junqueira Botelho et al.

Around this computer the first data processing centre was established (Rio Datacentro) at the Pontifical Catholic University of Rio de Janeiro (PUC Rio) and gave birth to the first Master’s programme in computer science as well as the first academic undergraduate course. This spurred the emergence of a software industry cluster which remained the country’s most advanced and largest until the early 1990s. However, three developments changed this picture over the next decade and half. First, the end of the failed nationalistic market reserve policy to develop an indigenous hardware industry in Brazil (Botelho and Tigre 2005), which had assisted the development of a few large associated software firms and public data processing and software development centres in Rio de Janeiro which then accounted for about 60 per cent of national production, eliminated a critical, albeit artificial, national government institutional support to the consolidating industry. Secondly, the consolidation of São Paulo as the country’s economic centre eventually led to the gradual departure of the banking and financial sectors from Rio de Janeiro to São Paulo, and depriving Rio de Janeiro of an ever-important demand source as the financial sector accounts for about one-third of the total domestic software demand (Botelho et al. 2005). Finally, the near completion of the transfer of government ministries and public agencies to the new capital of Brasilia inaugurated in 1960, including of large government data processing agencies and services, took away yet another significant market for software and related services, estimated at one-third of the total market (ibid.), at a time of government IT modernization and when it began to outsource.6 As a result of these, several of the largest software firms of the country based in Rio de Janeiro (although they were relatively small to middle-sized by international standards) drastically diminished in size or shut down and specialized software developers opened offices in São Paulo, eventually transferring their headquarters there. Yet, the large majority of experienced first-generation software developers who had worked for these companies set up one- to two-men firms to provide customized software development and related services (software maintenance, network management, web design, Internet provider, etc.). It is estimated that there are about 17,000 such software firms operating in Rio de Janeiro.7 A few large national capital firms that survived continued to expand, particularly in the training area, and large IT and software services foreign firms which originally had operations in the city (IBM, EDS, Unysis, Accenture, etc.) also continued to expand, although much of their growth took place in São Paulo.8 5.3.2

Industrial structure

A recent territorial-based study of the economic profile of the state of Rio de Janeiro commissioned by the Rio de Janeiro state , business federation (FIRJAN) and the state chapter of the government-regulated SMEs promotion agency (SEBRAE RJ) based on agglomeration economies, identified 15 such agglomerations including one out of three in IT, in the city of Rio de

Rio de Janeiro Software Cluster 181

Janeiro, by far the largest and the most significant one as the others are mainly satellite fiscal-based fictional clusters (Britto 2004; see also Constant 2007). Other significant SSI demand clusters in the metropolitan region are: telecommunications; petrochemicals, chemicals and plastic; and shipbuilding. Britto (2004), analysing social security data for the city of Rio de Janeiro (2001), identified 1,313 IT firms (all activities confounded) employing almost 20,000 people (see Table 5.1). The majority of the firms are micro and small with an average of 14.5 employees for the IT industry as a whole, with those with data processing activities exhibiting a larger number – 22.7 employees. The average wage per employee for the IT industry was R$1,970.00 in December 2001, rising for firms with IT systems consulting activities and other IT activities. In regard to employment distribution by firm size one finds the following features: a) employment concentration in smaller-sized firms with activities in IT systems consulting and equipment maintenance and repair; b) an even distribution across firm size in firms with activities in software development and other IT activities; c) concentration of employment in smaller-sized firms with activities in data processing. The study also identified 13 exporting IT firms, mainly first-time exporter micro-enterprises, which in 2002 grossed US$500,000 in hardware sales to the United States, Chile Colombia and Bolivia.

Table 5.1

Characteristics of the Rio de Janeiro City IT local productive arrangement

Statistical category (CNAE) – Integrated activities 72109 – IT systems consulting 72206 – Software development 72303 – Data processing 72508 – Maintenance and repair of office and IT equipment 72907 – Other IT activities, not previously specified Total

Wages Aver. size Employment # firms (12/2001 – R$) (employees)

Average wage (R$)

3,539

330

7,456,952.15

10.72

2,107.08

3,103

222

6,471,454.19

13.98

2,085.55

6,399

282

11,849,685.04

2.69

1,851.80

2,771

224

3,489,279.53

12.37

1,259.21

3,231

255

8,239,756.25

12.67

2,550.22

19,043

1313

37,507,127.16

14.50

1,969.60

Notes: CNAE means Classificação Nacional de Atividades Econômicas – it’s the name of the current national classification (Brazilian Activity Classifications) in accordance with the International Standard Industrial Classification. Source: Britto (2004).

182

Antonio José Junqueira Botelho et al.

5.3.3 Institutional evolution In 1965, the country’s first computer users’ society was created in Rio de Janeiro (Sucesu) and throughout the 1970s the city hosted the main IT events, Sucesu’s trade fair and congress. In the 1980s a new trade fair organization, Fenasoft, was created in Rio de Janeiro, but soon thereafter it is transferred to São Paulo, and its last edition in the city took place in 1996. In the following years business associations, promotion agencies and other organizations and institutions failed to act in a coordinated fashion to launch significant trade events in the city and, more importantly, failed to work together on a common project for the local SSI. From being the country’s IT capital, Rio de Janeiro gradually lost its role of industrial and cultural reference, although it remains the second largest regional market and industry. However, a reversal in this policy governance decline began in 2002 with the creation of the Rio Software Network, Redesoft, with an executive committee composed of several local institutions: SEBRAE RJ, Riosoft (the local chapter of the government funded software industry promotion organization SOFTEX), ASSESPRO RJ (Association of Information Technology, Software and Internet Firms), SEPRORJ (Data Processing Firms Business Association of Rio de Janeiro), Firjan, ACRJ (Chamber of Commerce of Rio de Janeiro), Rio de Janeiro State Government (Secretariats of Economic Development and of Science and Technology – Sede/Secti), Municipality of Rio de Janeiro, Reinc (Rio de Janeiro State Network of Incubators and Technological Parks) and Funpat (representing several partners of the Petropolis Technological Park, a nearby mountain town). In 2004, SEBRAE RJ commissioned a diagnostic study of the IT sector in the state of Rio de Janeiro to assist in the design of policies and mechanisms for the local SSI (SEBRAE RJ 2005). This joint governance effort and network operation eventually produced a collective project for the development of the IT sector called City of Rio de Janeiro IT Project, coordinated by a council formed by some of the same institution in the network (City of Rio de Janeiro Technological Network – REDETEC, SEPRORJ, ASSESPRO RJ, SEBRAE RJ) and the support of the State Economic Development Secretariat. The pioneering project in terms of a formal governance structure came to light in March 2005 and was made up of 19 structuring activity programmes, ranging from software development certification (CMMI),9 establishment of a an office to assist local firms in procuring public funds for financing software innovation and development to capacity-building in software engineering. It also set the following targets: (1) increase the volume of sales of IT firms by 10 per cent until December 2005 and 15 per cent until December 2006; and (2) increase the number of employees by 5 per cent until December 2005 and 10 per cent by December 2006. In September 2005, a SEBRAE RJ entry review of the Project found that the majority of the firms (93 per cent) had knowledge of the actions and 64 per cent felt it

Rio de Janeiro Software Cluster 183

benefited from them. A majority of the interview sample (52 per cent) participated in the decision-making process leading to the Project. Results indicators were to be monitored by subsequent reviews. By the end of the first half of 2008, the Project’s governance will have completed the first review of results against the targets, focusing on increase in sales and number of employees. The entry review numbers for the related indicators average sale volume per firm in the cluster ranged from R$5.06 million to R$10.84 million; and the average number of employees from 68 to 211. The great majority of sampled firms are members of the business trade organization (SEPRORJ), 76 per cent, and of the sector association ASSESPRO, 67 per cent. In the meantime, in 2003 the State Government of Rio de Janeiro through its economic development and science and technology secretariats launched the Responsible for the Programme to Promote Technological Knowledge in Information Technology of Rio de Janeiro State – Rio Knowledge, which has among its objectives to promote the development and the consolidation of the IT sector in the state and in the capital city. Recently, Constant (2007), identified three sub-clusters in the city: (1) Downtown Pole; (2) Rio de Janeiro Technological Park, at the Federal University of Rio de Janeiro (UFRJ); and (3) South Rio Pole, distributed according to Figure 5.1 and with the following characteristics: ●

●

●

Downtown Pole: Within an urban perimeter established by the Municipality of Rio de Janeiro of 7 square kilometres are housed about 54 per cent of the firms of the cluster, totalling over 500 firms with a significant concentration of software developers. In the same area are located the main organizations and institutions – ASSESPRO, SEPRORJ and Riosoft/Softex. Rio de Janeiro Technological Park profits from its proximity to Brazil’s largest federal research university and its graduate programme laboratories (Coppe/UFRJ), with the national Electric Energy Research Center (Cepel) and the corporate research centre of the state oil and gas company Petrobras (Cenpes), as well as a business incubator in the park area. South Rio Pole: Composed of firms located in the Gávea neighbourhood, it is in fact an externalization of the business incubation activities of PUCRio, which has the country’s highest graded computer science graduate programme according to the Ministry of Education graduate education promotion agency (CAPES), with a focus on distance learning, media convergence and software for engineering research products and processes developed at PUC Rio’s labs; and in neighbouring Barra da Tijuca, where the city’s largest firm is located, the state Cobra Computadores, a survivor from the market reserve policy; a Cinema and Video Production Pole and the Globo television network (Brazil’s largest) Production Centre (Projac). Several IT firms have began to move to this emerging sub-cluster thanks

184

Antonio José Junqueira Botelho et al.

to a state funded high-speed fibre optics network and it is estimated that about 18 per cent of the firms of Rio de Janeiro are located there.

5.3.4 Recent policy actions In light of the growth from 2006 of the number of public calls to support R&D and innovation in firms, particularly in software, one of the four industrial policy priorities (2004) and related areas such as digital TV through grants and low-interest loans, the Rio de Janeiro IT cluster institutional governance have launched efforts to diffuse these calls and to provide professional assistance in writing proposals. ASSESPRO RJ, a nonprofit professional business association representing a broadly defined local IT industry – IT, software and Internet – in September 2007 established a Project Management Office which assisted 40 interested associated local firms to prepare proposals to the second economic subsidy call of the Brazilian Innovation Agency FINEP. The call, with about US$260 million in resources, provided grants to corporate innovation projects in priority areas of the Industrial, Technological and Foreign Trade Policy (PITCE), including software and digital TV, areas which were allocated about US$60 million, a sharp increase from US$40 million in the previous 2006 call, when 114 projects in all areas received grants. The assistance package put together comprised a customized commentary on the call, a FAQ, hints for filling out forms and structuring projects and a set of recommendations. ASSESPRO RJ used the opportunity to generate a procedural discipline among its associated firms wishing to submit proposals in so far as it encouraged submission just from those willing to structure its proposal in a coherent and consistent way. It also learned lessons from previous calls and shared these with its associates. Some of the most common mistakes identified in project submission by associates in previous calls, mainly regarding project coherence that prevented them from getting to the merit analysis phase were: absence of technological innovation; lack of clarity in regard to objectives and methodology; ill-defined management coordination mechanisms; inadequate financial and project timetables; and lack of mention to call priority components (cooperation, sharing, result appropriation, etc.). In 2005, 23 firms from Rio de Janeiro received R$60.3 million (US$33.5 million) in government grants (FINEP, Faperj-State Foundation for Research Support and CNPq – National Research Council calls). In 2006, Rio de Janeiro IT firms had 16 projects approved in these public calls. ASSESPRO RJ PMO has also started to map associates’ demands in order to generate more focused assistance products. It has found that the five most sought-after types of resource are grants, venture capital, angel investment, counterpart grants and private equity, and that public resources are considered more attractive than private ones. Firms plan to use these

Rio de Janeiro Software Cluster 185

resources to fund business expansion, human resources, working capital, innovation and exports. In 2007, the state business federation FIRJAN published a short collection of position papers aimed at establishing a strategic framework for strengthening the software industry in the state (SISTEMA FIRJAN 2007). Though lacking in new data and analytical depth, the papers therein represents a small block in the ongoing construction of a renewed awareness of local industry, while also offering some policy recommendations. Finally, in July 2007, the Municipality of Rio de Janeiro submitted to the City Council a Programme of Incentives to Investments in the IT Sector in the City of Rio de Janeiro (Law Project 1250/2007). The programme’s main component is a proposal of reduction in the Service Tax (ISS) on SSI activities from 5 per cent to 2 per cent thus meeting a longstanding demand of the local SSI firms to bring the city service tax for SSI sales in line with that practiced by other major capitals (Porto Alegre) and towns in major metropolitan regions (Barueri and Hortolandia in the Greater São Paulo Metropolitan Region), what has been a major factor in the change of local firms’ headquarters to other cities, estimated at 50 per cent. Another is a series of fiscal incentives for SSI firm investing a minimum of R$50,000 (about US$30,000) comprising credits in the service tax (ISS) up to 80 per cent of investments, 50 per cent of real estate transfer tax (IBTI), and 50 per cent of property tax (IPTU).

5.4 Local development, innovation and the Rio de Janeiro software cluster This section aims at launching the basis for the discussion on the reformatting of local innovation clusters in developing countries. It presents a case study and an analysis of the current policy efforts to establish an innovation cluster in software in the metropolitan area of Rio de Janeiro. The motivations, mistakes and policy traps involved in this attempt for promoting local economic development and high-tech SMEs in Rio are put into relief. It argues that the lack of inclusion of local economic development in the policy recipes of the State of Rio de Janeiro, as well as a scant and weak local governance, favouring a purpose-driven institutional support are main factors. A coherent local governance capable of setting up a local economic development agenda designed to promote innovation in high-tech SMEs, thus capable of attracting the external resources required to add value to existing territorial assets as well as to developing new ones is missing. Without it, the high-calibre individual actors who do play a role in innovation will not be capable of agglomeration into a professional community to determine the pace of technological communication in local production systems of SMEs.

186

Antonio José Junqueira Botelho et al.

5.4.1 Rio de Janeiro local innovation system 5.4.1.1 Human resources According to the Brazilian Ministry for Education, there were less than 600 PhD titles granted from Brazilian universities in 1980. From 1996 to 2000, there were about 20,900 PhD titles awarded in the country. In 2004, this figure has reached almost 48,000 PhDs in diverse areas of scientific knowledge all over the country, although a significant amount of these titles have been granted from public universities in the richer south and south-eastern states. The majority of Master’s and Doctoral title holders work (2000) in the south-eastern states of São Paulo (22,354) and Rio de Janeiro (16,763). However, on a per capita basis we see that major cities like São Paulo and Rio de Janeiro, having together more than 23 million inhabitants, are being replaced in the top of the list by other municipalities, most of them with less than 1 million inhabitants, quality research universities with backgrounds in specific areas of knowledge, and a significant presence of qualified individuals holding Master’s and Doctoral degrees. Even though having dozens of universities with internationally celebrated academic and scientific achievements, both São Paulo and Rio de Janeiro suffer from chronic social disorders and growing urban violence that is causing a sort of brain drain to other cities offering more quality of life conditions, located nearby or even away from these urban centres. Yet, the cities of Rio de Janeiro and São Paulo are still the most important for the country’s GDP, and still account for the highest number of business incubators, science parks and start-ups based on potentially prospective new technologies. Education is another very important concern. The technical schools in the state of Rio de Janeiro, who prepare technicians with a relatively good quality for attending industry needs, received in 2005 only 11.5 per cent of the total enrolments in the state’s high-school system. The statistics for higher education – at university level – are even more skewed for both Rio de Janeiro and other Brazilian states. According to the Brazilian Ministry for Education, only 9 per cent of people from 18 to 24 years of age are enrolled in an undergraduate degree course. The State of Rio de Janeiro encompasses 13 per cent of the total enrolments in universities at undergraduate level, while São Paulo detains 26 per cent. That is to say, two out of 26 Brazilian states roughly encompass 40 per cent of all enrolments in undergraduate courses in Brazil. In Rio de Janeiro, in 2004, only 15.8 per cent of total enrolments corresponded to students in science and technological disciplines, against 19.1 per cent in the State of Santa Catarina and 44 per cent in South Korea. 5.4.1.2 Innovation infrastructure The city of Rio de Janeiro, although presenting a world-class R&D infrastructure, does not have a sufficient innovation infrastructure. Bridging the gap between a R&D and an innovation-oriented infrastructure will

Rio de Janeiro Software Cluster 187

fundamentally depend on coherent local governance capable of setting up a local economic development agenda designed to promoting innovation in the territory, thus capable of attracting the external resources required to add value to existing territorial assets as well as developing new ones. Without this, which is built by cognitive and normative collective goods, the highcalibre individual actors who do play a role for innovation to take place will not be capable of agglomeration around the professional communities that determine the pace of technological communication in local production systems of SMEs. The most salient aspect of the IT industry of Rio is the presence of top-level educational institutions in the state, plus a myriad of mature support organizations that are acquainted on the capital importance of the IT industry for the promotion of a sustainable development of the territory. The city of Rio de Janeiro is well positioned in terms of the capacity for knowledge production, in respect to other major Brazilian cities. It hosts a high number of holders of Master’s and PhD titles, and is well positioned in terms of research centres and research universities in diverse areas of knowledge. In areas of scientific knowledge directly or indirectly related to support institutions and universities, only to mention those in the metropolitan area of Rio de Janeiro, there are seven main celebrated research universities: the Federal University of Rio de Janeiro (UFRJ); the Federal Centre for Technological Education (CEFET); the Catholic University of Rio de Janeiro (PUC-Rio); the State University of Rio de Janeiro (UERJ); the Institute for Applied and Pure Mathematics (IMPA); the Federal Fluminense University (UFF); and the Military Institute of Engineering (IME). Apart from IMPA and IME, all of these universities have business incubators hosting firms operating in the IT sectors. All of them offer Master’s and PhD courses on areas pertaining to IT, and the average distance from one university to another is usually less than 20 kilometres. In addition, there are also another public university in humanities (UniRio) and more than 40 private universities offering undergraduate courses and vocational training in IT areas, as well as hundreds of technical schools offering vocational courses for the qualification of computer programmers, electronic and telecommunications technicians and so forth. There are also business incubators not related to universities, as the ones hosted by Instituto Nacional de Tecnologia (INT) and Serviço Nacional de Aprendizagem Comercial (SENAC-RJ). The city of Rio de Janeiro counts with several support institutions. The most important of them is SEBRAE-RJ. SEBRAE-RJ designs and operates entrepreneurship development programmes in many commercial, services and industrial areas, and is structuring new lines of action for more effectively stimulating the formation of local production systems of SMEs in technological fields. The recognition on the part of SEBRAE of the importance of clusters of firms for local economic development, with the translation of support measures for those firms, started in the beginning of 2000 and has been strengthened after the joint project carried out with Promos (see note 94). In addition

188

Antonio José Junqueira Botelho et al.

there are, located in the same building in downtown Rio de Janeiro: the export promotion agency for IT firms, RioSoft; the Association of IT firms, ASSESPRO; and the Union of Workers of IT firms of the State of Rio de Janeiro, Sindpro. In addition to these capillary support institutions, given that most of them have existed since the late 1970s, the city of Rio hosts the headquarters of the most important science and technology and industrial development federal institutions, namely the Brazilian Innovation Agency (FINEP) and the Brazilian Development Bank (BNDES). These institutions, spread in the city of Rio de Janeiro, are located an average of 10 kilometres away from one another. Even hosting so many institutions, the level of utilization of centres of excellence from small and medium-sized enterprises is very poor. The penetration of development support organizations is also considered by firms and policymakers alike as embryonic, given that the critical mass formed around the importance of systemic institutional support for high-tech SMEs take-up has started to be included in the policy arena a little more than five years ago, with very few substantial practical results. In some cases, research centres and firms compete for the scant public funds in calls and tenders for the financing of technological projects on the part of FINEP. According to Rede de Tecnologia do Rio de Janeiro (hereafter Redetec), the Rio de Janeiro network of innovation support institutions to start-ups, there are 19 business incubators in the State of Rio, with 80 per cent of them concentrated in the city of Rio and the metropolitan area (the city’s surrounding municipalities). These business incubators, 90 per cent of them formed within or nearby important universities, support high-tech start-ups on the most critical phase of their growth stages, that is the pre-incubation (business planning support, in most cases) and incubation processes (physical and telecom infrastructure provision, managerial and technological coaching). The sectors covered are so vast, in many cases exploring the core research competences of local universities, ranging from biotechnologies, IT, agro-industry and so forth. In 2006 there were, according to Redetec, 107 firms being ‘incubated’ in the state of Rio, as well as some more 103 firms that managed to survive to the incubation process – which takes two to three years – thence established in the marketplace. 5.4.1.3

IT industry structure and markets

A sustained exodus of firms is being verified in the city of Rio de Janeiro, fundamentally motivated by the problems related to public security. A survey carried out by the Federation of Industries of the State of Rio de Janeiro (hereafter FIRJAN) in 2003 with 2,665 workers in the city of Rio, concluded that 44 per cent of those people declared to have been victims of any type of violence caused by third parties. Moreover, in another poll carried out by FIRJAN in 2005 with 1,157 workers, 70 per cent of these interviewees named public security as a priority problem to be sorted out in the city of Rio. SMEs

Rio de Janeiro Software Cluster 189

are an important engine of the Brazilian economy, representing, according to SEBRAE (2004), more than 99 per cent of the total of firms legally established; Brazilian SMEs employ 44 per cent of the total workforce and account for 20 per cent of the country’s GDP. In the state of Rio de Janeiro, SMEs represent 26 per cent of the total of legally established firms in the country. In the state of Rio de Janeiro, the level of exports of SMEs is inferior to the national average. SMEs established in Rio de Janeiro managed to export only 6.9 per cent of the total exports in the state. If the medium-sized enterprises, say those employing from 50 to 100 people, are excluded from these statistics, small and very small (micro) firms account for less than 1 per cent of the total exports. These results are indicative of a paradox pertaining to a booming telecommunications and energy economy that is also the second GDP in the country, responding to 12.8 per cent of the Brazilian GDP. If Rio de Janeiro was an independent country, it would have been the sixth largest Latin American economy (FIRJAN 2006). This apparent socioeconomic paradox has important implications for the establishment of local production systems of SMEs in Rio de Janeiro. Local Production Systems of SMEs in Rio de Janeiro are more strongly spread in the metropolitan region (the capital city and surrounding municipalities) and in parts of the south and south-east of the state, given the proximity of the latter to ‘Porto de Sepetiba’ (the main port in the state) and to the state of São Paulo, the larger consumer market in Latin America. Local production systems are growing in importance in Rio, although not much is yet known in terms of the real strength of SMEs to promoting quality cooperation, for creating quality jobs and for attracting long-term private and public external investments. The main motivation behind this reasoning can be derived from the low presence of local production systems formed by SMEs in the state. The most relevant industrial segments in Rio are, on the one hand, those of commodities, like the production system of petroleum and renewable energy formed around the northern coastal city of Macaé, and strongly dominated by Petrobras – the Brazilian state-owned oil giant – and major international oil companies. In addition, there are those local production systems in the auto-parts segment, formed in the south-eastern cities of Resende and Porto Real, particularly around the facilities of Volkswagen and Peugeot-Citröen, who have been attracted less than a decade ago by higher fiscal incentives offered by the state government. The same applies to the petrochemicals local production system, located on the edge of the metropolitan area of the city, and also strongly dominated by large firms, as well as the shipping local production system in Niterói and the steel industry cluster of Vale do Paraíba region. In these systems, small and mediumsized firms are mostly embedded in hierarchical networks of suppliers or by outsourcing parts of the production processes of larger firms (Avgerou and La Rovere 2003; Pinto 1999). That is to say, the existence of larger firms

190

Antonio José Junqueira Botelho et al.

established in those municipalities conditions the attractiveness of firms of more reduced dimensions, as well as suppliers of services, private and public universities and schools. The innovative potential of these smaller firms tends to be conditioned by the technological development strategies of those larger firms dominating the respective production chains of these local production systems. Technological innovations in SMEs operating in these industrial segments, whenever there are, tend to be more concentrated on the improvement of existing industrial processes (process innovations) rather than on radical product innovations that could have promoted disruptive changes in the production chains to which these firms belong. This is indeed a characteristic of the Brazilian industry whose innovation is much below international standards (De Negri and Salerno 2005). In the other sectors, the production process is much more concentrated on SMEs, and the system’s governance is thereby less influenced by larger firms. However, apart from the IT and media production sectors, those industries are mostly formed around low-tech and highly labour-intensive sectors. The mapping of local production systems in Rio, as the ones already carried out with the institutional support of capillary institutions as SEBRAE RJ and FIRJAN, although very important for stimulating a debate on more targeted public policies to stimulate the growth of local production systems, has so far neglected the different character of those SMEs willing to operate on the edge of existing technological prowess. The needs of these firms and the characteristics of these production systems, in terms of collective goods and local governance specifications, imply different policy designs and institutional support, as exhaustively discussed in this research work. The IT sector is no different reality in Rio de Janeiro – and in Brazil as well – even though it is among the fastest growing sectors of the Brazilian economy. The strategic importance of mature local production systems in industries that demand significant IT products and services is somehow substituting incipient public support, in that sense indirectly stimulating the development of this sector in Rio de Janeiro. Nevertheless, the higher value-added demand of those industries for IT solutions manages to cover only the most competitive IT firms in Rio de Janeiro, which are usually those of medium and larger dimensions. In addition, these more dynamic sectors can manage to cover only a fraction of the potential demand of other important sectors who demand IT products and services from local firms. According to Riosoft, an export promotion agency for IT firms originating in Rio de Janeiro, the market for IT firms in Rio is divided as follows: 40 per cent public administration; 40 per cent financial sectors; 20 per cent other sectors (ASSESPRO 2005). The major chains – of international relevance – hosting important local production systems in the state of Rio de Janeiro, such as shipping, oil and renewable energy and tourism evidence a more dynamic expansion of certain layers of the IT sector in Rio de Janeiro, thereby presenting interesting

Rio de Janeiro Software Cluster 191

peculiarities in the demand for sophisticated IT solutions. Hence, the dynamics of the demand for IT products and services in Rio de Janeiro is directly dependent on the expansion of major production chains and to the purchasing power of the public apparatus. SMEs are also considered as important sources for IT products and services, although the sophistication of IT solutions targeting SMEs tends to be much more modest, for most of the SMEs in the state of Rio de Janeiro pertain to traditional sectors of the economy as well as to services-related areas (Avgerou and La Rovere 2003). On the supply side, though, there can be seen a major concentration on the development, adaptation and customization of products and services to SMEs, rather than onto higher value-added customers who demand more sophisticated solutions from IT firms. According to the Brazilian SME Support Agency – SEBRAE, the market distribution of IT solutions originating from SMEs is fundamentally concentrated onto other SMEs. Only a very few IT SMEs are capable to export their outputs, most of them exploring foreign market niches in other Latin American and in Portuguese-speaking African countries. Those SMEs exploring potentially new markets are solely the start-ups originating in business incubators and in science parks. The provision of IT solutions in Rio de Janeiro is much concentrated on SMEs. According to the Association for Information Technology Firms in the State of Rio de Janeiro-ASSESPRO data from 2004, 94 per cent of IT firms in Rio de Janeiro have a turnover below R$5 million a year (or US$2.9 million), and 84 per cent of them earn less than R$2 million a year. In addition, according to the same survey, 90 per cent of IT firms have fewer than ten employees. The national reality is no different, for 53.1 per cent of Brazilian IT firms earn less than R$1 million a year, according to the Brazilian Software Promotion Agency (SOFTEX). According to ASSESPRO, the market segments covered by IT firms from Rio de Janeiro concentrate on vocational training, technical support, sales and Internet access provision. Although presenting one of the most important IT industries in the country, hosting national leaders in the development of IT solutions in certain sectors, as those specifically targeting, for example, the oil and tourism industries, the IT industry in the State of Rio is losing position relative to other regions. The state of Rio Grande do Sul, for example, presented a sustainable 11 per cent average growth rate from 1998 to 2004 (ASSESPRO 2005). There are no official – that is to say, agreed by all – statistics on the real number of IT firms in the State of Rio de Janeiro. The Brazilian Bureau of Statistics-IBGE, the Ministry of Science and Technology-MCT, Riosoft, SEBRAE and ASSESPRO have each of them different methodologies and hence reach different indicators. The number usually agreed upon as the most approximate is the one offered by ASSESPRO, on the basis of its National Register of Enterprises of Information Technology Services,

192

Antonio José Junqueira Botelho et al.

indicating that, in 2006, there were almost 10,000 firms in the IT business in the state of Rio. The question that intrigues businesspeople and policymakers alike tends to be then geared towards understanding why, despite having capillary institutions, top research centres and so many firms of different sizes and technological capacity-building statuses, the level of institutional and interfirm cooperation is practically non-existent, and consequently technological innovation among high-tech firms in Rio de Janeiro is so modest. There are public financing programmes, some of them old, some others new; there are private equity investors interested in new investment opportunities; there is a community of business angel investors looking for prospective investment opportunities; there are business incubators and science parks covering the most relevant areas of scientific knowledge with prospective business development potential; there are some legal attorneys prepared for dealing with complex intellectual property issues. Albeit with varying degrees of institutional maturity, the presence of these agents, most of them object of past industrial policies, tend to target the variables that hamper the competitiveness of existing firms, in particular those in the traditional sectors of the economy. Even those programmes specifically targeted to high-technology firms, such as the Sector Funds of FINEP, the tax exemptions from the state and city governments, and the export-oriented policies and so forth have been designed for individual firms themselves. The collective nature of knowledge continues to be neglected. The reason for this is the persistent institutional refusal of local economic development as the driver for the building up of the local collective goods needed to strengthening the innovation potential of established and new high-tech firms. 5.4.1.4 Collective goods: How important for software firms in LIS? Although recognized as an important driver of territorial development in Rio de Janeiro (ASSESPRO 2005; SEBRAE 2004), the growth of the SSI industry rooted in SMEs is facing significant barriers towards a more sustainable development. In contrast to other technologies like bio and nanotechs, which fundamentally target potentially new markets for their products, most ITs – software in particular – are mature technologies whose development is consistently dependent on the demand side. That is to say, the more sophisticated the demand for IT solutions is, more high-tech-based products and services in IT could have been developed so that firms would naturally tend to engage in complex networks for being able to meet those more sophisticated demands of customers. In terms of innovation, the role of government tends to be more limited than other channels, as could have been the market itself and the strengthening of R&D in universities and firms. The complex, costly and long-term nature of innovation limits the potential of governments as an inducer of technological innovation in firms by means of its purchasing capacity. The overall disbursements of the

Rio de Janeiro Software Cluster 193

government of the state of Rio de Janeiro in science and technology (S&T) research has not reached R$140 million a year (or US$82 million a year) in 2005. On the other hand, in the state of São Paulo in that same year, the S&T disbursements of the government of that state was of R$1.5 billion a year (about US$900 million). In part this discrepancy is explained by the fact that Rio de Janeiro has more federal institutions than São Paulo, so that Rio takes advantage of federal investments in R&D science and technology research. The most important R&D institutions in Rio de Janeiro are federal, not state ones. The differences in terms of infrastructure for research, programmes for the qualification of researchers and scientists and the quality of the workforce trained among federal and state institutions is remarkably significant. There are many more federal universities and R&D labs spread throughout the country, so that Rio increasingly hosts just a fraction of them. Furthermore, federal investments in education have been decreasing in recent years, and there is a growing tendency for these funds to become even more. If state governments, as that of Rio de Janeiro, do not engage in supporting more R&D activities, the most likely outcome will be translated as a significant loss of R&D disbursements in Rio de Janeiro in respect to other Brazilian states. Without a proper R&D base working together with firms in complex projects, there cannot be developed the set of innovation-oriented competencies that are in line with domestic and international demands for advanced and intensive knowledge-based products. As a consequence, most R&D institutions in Rio de Janeiro suffer from lack of scale for long-term R&D and brain losses to other Brazilian states and countries (particularly the US). In addition to these institutional deficiencies, there is a ‘project seeking’ culture within universities and research institutes – in particular among young researchers – thus transforming these institutions in service support organizations to industry, rather than into true partners to strengthen the technological capabilities of the private sector. University– industry cooperation is acknowledged today as a pillar of technological innovation, given the externalities arising out of such cooperation in terms of learning opportunities, shared resources and competences put together for achieving common objectives. The local presence of science parks and business incubators is not a determinant for the innovative success of high-tech firms. However, the recent growth of these is indicative of the growing awareness of the need to bring together firms and the centres of knowledge production so as to increase the potential of researchers and entrepreneurs to engage in potentially innovative scientific projects as well as in start-up creation. As international evidence suggests, innovation tends to take place more easily when there is a sustainable combination of domestic and international market demand as well as a local R&D base sustaining the formation of quality workers, researchers and scientists together with quality research

194

Antonio José Junqueira Botelho et al.

with business development potential. In both these aspects – market and R&D structure – the state of Rio de Janeiro is not well positioned to spur on IT-based innovations. International market demand for advanced IT products and services is being addressed by more competitive Indian, Taiwanese, Israeli and US firms (Botelho 2005; Bresnahan et al. 2001; Saxenian 2006). The domestic market is limited and conditioned, in respect to international standards, by the more limited investments of Brazilian firms in the acquisition of capital goods (hardware and software comprised) and in the development of new products, in that sense hampering the internal market development potential for higher value-added IT solutions. In the cases of larger firms – national or multinational ones – the demand for ITs is usually met by larger foreign and Brazilian IT firms, leaving a very limited room for SMEs and start-ups to get into these markets. On the R&D side, the situation is not different. The evaluation on the impact of R&D in the ICT sector in Rio de Janeiro and other Brazilian cities has not yet been determined, although the importance of R&D to stimulate the development of an IT industry together with spin-offs and start-ups from local universities and larger firms cannot be ignored, being thereby sufficiently explored in the international literature (Saxenian 1994; Shavinina 2003; Storper 1997; Swann et al. 1998). In Rio de Janeiro, institutions of a corporate, governmental and support nature coordinate a number of activities dedicated to the IT sector. The state government of Rio de Janeiro has an IT development programme named ‘Rio Conhecimento’, giving fiscal incentives to firms in that sector. The highest investor in support programmes is SEBRAE RJ, which invests more in support programmes for the IT sector in Rio than all the other institutions together (including BNDES and FINEP). A recent survey, carried out by ASSESPRO in 2004, concluded that Rio lacks an institutional articulation of agencies, projects and programmes. Actually, no general guidelines, objectives and development strategies are shared, in that sense hampering an effective mobilization of existing instruments, as well as legitimating new ones. The state of Rio de Janeiro has most of the critical institutions to reformat an innovation-oriented IT cluster. Nevertheless, there has never been established a thorough local economic development strategy in line with its institutional development potential. The complex socioeconomic situation partly conditions the growth of high-tech firms, thereby impacting the overall technological innovation potential of the state. This scenario demands an equally complex solution in order to, at the same time, improve the quality of jobs, quality of life of individuals and families, social inclusion and social security. While lives of innocent people keep being lost in the constant confrontations among policemen and criminals in the middle of crowded streets of Rio; and while the culture of tension and fear takes possession of the lives of citizens, the city will continue to be afflicted in its capacity to attract qualifying investments, thereby being powerless to

Rio de Janeiro Software Cluster 195

qualify its existing local collective goods and to develop new ones. Among the most important of the city’s assets are individuals with the business, scientific, legal and other competencies considered important for innovation to take place. But those people are not engaged in professional communities so that cooperation among them is neither strengthened nor encouraged. In light of the above, it is proposed that there is a missing ‘bridge’ between R&D and technological innovation, which are different things per se, and which must be established through the building up of local collective goods. These collective goods, considering the complex framework of the city of Rio de Janeiro, may be built by a local governance capable of setting up a broad and encompassing local economic development agenda. In order to understand the relevance of collective goods for high-tech SMEs, a questionnaire with a list of desirable local collective goods (Table 5.2) was passed to a sample of 11 IT start-ups which underwent incubation on the premises of local research universities. They were asked to rate their importance and pattern of use in the course of their evolution. Results showed that 77 per cent of the entrepreneurs interviewed declared they had used some sort of mentoring from experienced professionals in the marketplace; all of them use (and they do need, obviously) telecom infrastructures in their daily professional activities; all of them have made use of the infrastructure provided by business incubators and science parks in universities and research centres. Moreover, 66 per cent of them had made use of some sort of public seed capital, against only 10 per cent who claimed to have been invested in by venture capitalists and another 30 per cent who declared to have received funding from business angels. Coaching from academics, usually from former professors and supervisors in research projects, have been rated 4 (i.e., considered fundamentally important for them) by only 30 per cent of the entrepreneurs who admitted to have used this in the lifecycle of their start-up companies. When asked whether they consider they have the required knowledge to create new products and services, all of the entrepreneurs interviewed declared to possess such competences, thereby indicating the strategic importance of cooperation for technological innovation in start-ups is not considered relevant for them. This view can be derived by the existing gaps in entrepreneurship educational programmes in Brazil, where cases of best practice derived from real experiences of technological innovations that have become possible due to cooperation tends to be more limited and not treated on a primary perspective. Structured entrepreneurship educational programmes in Brazilian universities are practically non-existent. In Rio de Janeiro, the only comprehensive programme of this kind is provided by the business incubator of the Catholic University of Rio de Janeiro (PUC-Rio). This programme is geared at undergraduate students from any knowledge discipline of that university – technological, scientific and even humanities – who are willing to undertake a start-up, whether or not in the premises of PUC-Rio’s business incubator. This educational

Antonio José Junqueira Botelho et al.

Technological

Table 5.2

Local collective goods for the City of Rio de Janeiro

• Business angels’ communities. • Business incubators and science parks. • Coaching from academics in universities. • Management consulting companies (marketing, cost engineering, etc.). • Mentoring from experienced market specialists. • Public and private universities strongly engaged in R&D related to ICT sectors. • Public support to technological capacity programmes in start-ups. • R&D laboratories of large firms. • Seed capital (public and private). • Technological consulting companies (business plans, technology selection, engineering design, etc.). • Telecommunications infrastructure. • Venture capital (public and private).

Institutional

196

• Technical schools. • Technological capacity in public administration to improve the quality and extension of its services. • Traditional financial system (banks, credit unions, etc.). • Territorial marketing promotion on the part of the public sector. • Introduction of competencies for increasing the management skills of bureaucrats (project management) in large projects aimed at benefiting the territory. • Entrepreneurship educational programmes at schools and universities. • Centres of excellence in higher education (undergraduate and graduate courses as well as advanced vocational training). • Transport infrastructure (airports, roads, etc.).

Territorial • Leisure infrastructure (parks, green areas, etc.). • More reduced levels of corruption in the public administration at local, regional and national spheres. • Public health system. • Public security (quality, efficient and less corruptive police, lower levels of criminality). • Rule of law (or a legal system that fastens judicial processes, as well as that guarantees the fulfilling of the terms of business and civil contracts). Source: Alves (2007).

programme offers 11 courses ranging from entrepreneurial finance to marketing, from business planning to accounting, from notions of psychology for entrepreneurs to oral presentation techniques, and so on. About 600 undergraduate students attend these courses every year at PUC-Rio. All of the entrepreneurs from PUC-Rio’s business incubator have passed through one or more of these courses. Nevertheless, the main feature of entrepreneurship educational programmes in Brazil is that these courses are designed to be, at the same time,

Rio de Janeiro Software Cluster 197

both practical and in an accessible language to its target public, usually consisting in young undergraduate students. In most cases, these programmes include theoretical discussions of the phenomena of technological innovation and entrepreneurship, in that sense concentrating in themes with a more firm-oriented (individual) nature – like business planning, finance, accounting, marketing, etc. – rather than on incentives for the firms to work on jointly since their very start, as may happen in other international contexts. These are indeed part of a territorial development strategy, given that initiatives like technological entrepreneurship educational programmes need to be designed with the support of qualified local actors who need to be acquainted on the importance of this as well as other issues for establishing a functional framework at territorial level to strengthen the innovative potential of those risky ventures of more reduced dimensions. In terms of institutional collective goods – that is to say, those local collective goods whose presence in a given territory enhances the capabilities of local institutions of fulfilling their territorial missions – there have been considered as more important ones: transport infrastructure (rated 3.25); universities (rated 3); technological educational programmes (rated 2.75), and the project management competences in the public bodies (rated 2.75). Given that this survey concentrated on firms graduated from business incubators, it is quite natural that these firms point out universities as a very important local collective good. Local collective goods whose externalities produce conditions of the quality of life of citizens have a more apparent indirect impact in the activities of start-up companies. It can be noted that the rule of law was considered as critical (rated 4 by all interviewees). Among the entrepreneurs interviewed, 70 per cent of them declared their businesses had been affected by the slowness and lack of clear rules in the Brazilian judiciary system. The way in which the Brazilian legal system is structured, according to most of the entrepreneurs interviewed, does not enhance cooperation among firms for their fear that intellectual property rights and terms of contracts are not properly secured. The other aspects that condition the quality of life of citizens have been rated, on average, as such: public security, rated 3.25; corruption, 3.25; leisure infrastructure (squares, parks, green areas, etc.), rated 2.75; and public health, rated 2.25. Matters like public security and corruption have been considered as impacting the business of the interviewees in a very important manner.

5.5

Conclusions

While issues of a socioeconomic nature are not properly treated, the capacity of the city of Rio de Janeiro to attract qualified professionals and investment in the large scale will not come about as spontaneously and gradually as it did for the remote industrial policies of the past, even though the software

198

Antonio José Junqueira Botelho et al.

service clusters in India are also plagued by poverty-related issues. But, until quite recently, economic opportunities for professionals in India were more limited than in Brazil. Moreover, different times require different rules, and different rules, different policies. Without a local economic development strategy capable of bringing to life (or even of reformatting some of) these technological, institutional and territorial collective goods, it will be very difficult to place the city of Rio de Janeiro in the same high-tech route as other cities in developing countries. Technological innovation in new uncertain sectors – such as some areas in IT such as software and services – is also a by-product of social inclusion. The ability to produce the right mix of local collective goods becomes crucial today, not only to an individual area but also to an overall national economy. However, local collective goods favouring new knowledge creation can be more efficiently built by local institutions rather than by national policies. The latter should have a coordination role, in particular for those collective goods whose externalities produced in a territory directly condition the development of high-tech companies in local production systems of SMEs. In the prevailing cluster policy-making vision, similar policy recipes tackle similar problems. The foundations of innovation taking place in rather different local production contexts are interpreted as the same production phenomena, in a similar way to a prototypical view of the Italian industrial district experience. It is indeed a consequence of the novel character of the theme in Brazil, in which it has looked upon previous successful international experiences in order to identify interpretative models for the Brazilian case. In this view, the Marshallian industrial district experience has proved providential to an extent. That is to say, the acceptance of the role of external economies as both the result of cooperation and the motivation for the attractiveness of qualified professional newcomers and the overall recognition on the importance of collective goods as the hidden elements behind the systems’ performance. However, it has not been considered in the Brazilian public policy debates that collective goods need to have different natures in order to produce those external economies required for innovation to take place. The main remaining question about the transformations of industrial agglomerations or local productive arrangements in the context of developing countries is how to convert these inherently static firm agglomerations into evolving dynamic local innovation clusters. The European Union started to pursue a similar strategy a long time ago, without significant results in terms of innovations created in respect to what has been invested so far. Such policy debate is still under way here as there, but it must be recognized in both places that at least a lot has been done to understand the specific characteristics of innovation-based local production systems of SMEs. Rio de Janeiro is no exception and its still unfulfilled experience of reformatting an innovation-led software cluster exemplifies the scale

Rio de Janeiro Software Cluster 199

and scope of the challenge. The evaluation of potentially innovative ideas requires skills, competences and priorities that cannot be found in the public administration. Making territories capable of building the cognitive and normative collective goods required to have the right people in for innovation to take place demands long-term commitment, responsibility, leadership and engagement from all-important territorial actors. This chapter has contributed to bringing out additional elements and boundary conditions and reinforcing the role of existing ones in a modified flowchart model for an innovation-driven service industries cluster. Transforming a service industry agglomeration into a true LIS in SSI, which aims to compete on factors other than labour costs in a rapidly evolving and expanding complex market, requires innovation-driven demand and pragmatic collaboration (Helper et al. 2000) among firms with learningby-monitoring; and large firms as anchors, a necessary but not sufficient condition. For, increasingly, large anchor firms in the SSI industry, either domestic or multinational, are beginning to realize that by growing small firms and promoting their joint growth into a innovation-oriented cluster is also a means for them to compete internationally beyond labour costs due to the following emerging trends: (1) the scope of offshore outsourcing is increasing and is deepening with the rise of business process outsourcing; (2) SMEs software and services demand is growing in size and sophistication thus constituting an increasingly important market for the SSI industry; and (3) the growth in increased market fragmentation demands from software developers and service providers gives a different level of flexibility associated with diverse expertise and continuous quality enhancements.

Notes 1. In the mid-1980s IBM Brasil had just 2,200 direct employees versus 7,200 in 2007, completed by 3,000 indirect posts. Its software services growth in Brazil has been spectacular and in 2005 over 1,000 programmers were hired at the Hortolândia Centre, which is expected to have 10,000 employees by 2010. 2. For comparison, at the end of 2007, EDS employed 38,000 people in India, expected to reach 25,000 at the end of 2008. 3. This section builds upon the discussion of industrial districts in Alves (2007). 4. It must be further noted that the idea of a compact social structure does not imply an absence of clashes of interests and conflicts and that a local production system, as defined, can be characterized by a social situation which is highly disintegrated, or even by the presence of a booming illegal economy, as in certain southern Italian regions. The latter aspect is also a very present – and worrisome – issue in the slums of Rio de Janeiro in Brazil, where drug lords hinder local economic development by dictating rules and roles to be followed, affecting, in a substantial way, the lives of most of their inhabitants and, more importantly, of the city as a whole. 5. There is great divergence in studies about the number of firms in the SSI: foreign consultants say 10,000, ASSESPRO counts 27,000, of which 4,200 are in software

200

6. 7.

8.

9.

Antonio José Junqueira Botelho et al.

development and the Brazilian statistical office (Instituto Brasileiro de Geografia e Estatística IBGE) estimates 38,000 IT firms employing 220,000 people. This spurred over the next few years the creation of the data processing services industry cluster in Brazil. However, several of these micro software firms have fiscal headquarters in neighbouring counties with lower services taxes, which represent the major financial charge on their business activity. Over the past decade new poles have emerged from the southernmost state capital city of Porto Alegre to the north-east economic capital of the state of Pernambuco, Recife as well as in more recent years increasingly in the interior middle-sized cities of the state of São Paulo (e.g., Sorocaba), in the São Paulo city neighbouring towns of Campinas and Barueri, and in the reconverted declining industrial areas of the so-called Greater São Paulo ABC (municipalities of Santo André, São Bernardo and São Caetano) region, the hub of Brazil’s automotive and manufacturing industries. According to ‘Software Engineering Institute (SEI)’ CMMI or Capability Maturity Model® Integration (CMMI) is a process improvement approach that provides organizations with the essential elements of effective processes. It can be used to guide process improvement across a project, a division or an entire organization. CMMI helps integrate traditionally separate organizational functions, set process improvement goals and priorities, provide guidance for quality processes, and provide a point of reference for appraising current processes.

References ABES (2007) Mercado Brasileiro de Software: panorama e tendências. São Paulo: Associação Brasileira das Empresas de Software. Alves, A. and da S. (2007) Local Economic Development and Innovation in the Information Society – A Comparative ‘Political Economy’ Assessment of the Role of Collective Goods for Innovation in Brazilian Technological Districts. Tesi di dottorato, Università Degli Studi Di Milano-Bicocca, Progetto QUA_SI, Corso di Dottorato in Società dell’Informazione, Milano, Italy. Amendola, M., C. Antonelli, and C. Trigilia (2005) Per lo sviluppo: processi innovativi e contesti territoriali. Bologna: Il Mulino. Antonelli, C. (1986) L’attività innovativa in un distretto tecnologico. Turin: Edizione della Fondazione Agnelli. —— (2001) The Microeconomics of Technological Systems. Oxford: Oxford University Press. Associação Brasileira Das Empresas De Tecnologia Da Informação, Internet E Software (ASSESPRO) (2005) Levantamento da Indústria de TI no Estado do Rio de Janeiro (2004). Rio de Janeiro. Avgerou, C. and R. L. La Rovere (2003) Information Systems and the Economics of Innovation. Cheltenham: Edward Elgar. Bagnasco, A. and Sabel, C. (1995) Small and Medum-Size Enterprises. London: Pinter. Becattini, G. (1979) ‘Del “settore” industriale al “distretto” industriale. Alcune considerazioni sull’unità di indagine dell’economia industriale.’ Rivista di Economia e Politica Industriale, No. 1, pp. 30–63. Becattini, G. (2000) Il distretto industriale. Un nuovo modo di interpretare il cambiamento economico. Torino: Rosenberg & Selier.

Rio de Janeiro Software Cluster 201 Botelho, A. J. and P. B. Tigre (2005) Country Report Brazil, Research Project ‘Comparative Study on East Asian and Latin American Information Technology (IT) Industries. Santiago: CEPAL. Botelho, A., G. Stefanuto, and F. Veloso (2005) ‘The Brazilian Software Industry,’ in From Underdogs to Tigers: The Rise and Growth of the Software Industry in Some Emerging Economies, A. Arora and A. Gambardella Eds, Oxford: Oxford University Press, pp. 99–130. Bresnahan, T., A. Gambardella, and A. Saxenian (2001) ‘ “Old Economy” Inputs for “New Economy” Outcomes: Cluster Formation in the New Silicon Valley.’ Industrial and Corporate Change, Vol. 10, No. 4, pp. 835–860. Britto, J. (2004) Arranjos Produtivos Locais: Perfil das Concentrações de Atividades Econômicas no Estado do Rio de Janeiro. Rio de Janeiro: SEBRAE/RJ, pp. 241. Brusco, S. (1992) ‘The Emilian Model: Productive Decentralisation and Social Integration.’ Cambridge Journal of Economics, Vol. 6, pp. 167–180. Camagni, R. (1999) Innovation Networks: Spatial Perspectives. New York: John Willey and Sons. Castells, M. (2000) The Rise of the Network Society (2nd edn). Oxford: Blackwell. Constant, I. (2007) Análise e Projeto de Desenvolvimento para o APL de Tecnologia da Informação da Cidade do Rio de Janeiro – Curso de Gestão de Projetos em APLs. Brasília: Comissão Econômica para América Latina e Caribe (CEPAL). Cooke, Ph. (2001) ‘Regional Innovation Systems, Clusters and the Knowledge Economy.’ Industrial and Corporate Change, Vol. 10, No. 4, pp. 945–974. Crouch, C., P. Le Gales, C. Trigilia, and H. Voelzkow (2004) Changing Governance of Local Economies: Responses of European Local Production Systems. Oxford: Oxford University Press. De Negri, J. A. and M. S. Salerno (2005) Inovações, padrões tecnológicos e desempenho das firmas industrias brasileiras. Instituto de Pesquisa Econômica Aplicada (IPEA), Brasília. DeVol, R. and A. Bedroussian (2006) Mind to Market: A Global Analysis of University Biotechnology Transfer and Commercialization. Santa Monica, CA: Milken Institute, September. FIRJAN (Federação das Indústrias do Estado do Rio de Janeiro) (2006) Mapa do Desenvolvimeno do Estado do Rio de Janeiro: 2006–2015. Rio de Janeiro: Firjan. Granovetter, M., E. Castilla, H. Hwang, and E. Granovetter (2000) ‘Social Networks in Silicon Valley,’ in The Silicon Valley Edge, Ch. Lee, W. Miller, M. Gong Hancock, and H. Rowen Eds. Stanford: Stanford University Press, pp. 218–247. Helper, S., J. P. MacDuffie, and C. F. Sabel (2000) ‘Pragmatic Collaborations: Advancing Knowledge While Controlling Opportunism.’ Industrial and Corporate Change, Vol. 9, No. 3, pp. 443–483. Kabir, J., D. Eaton, and K. Yoshida (2007) ‘The Role of Innovation in Building a Sustainable Technology Cluster: The Austin Case’, in The Flowchart Approach to the Formation of Industrial Cluster: Focusing on the Endogeneous R&D and Innovation Mechanism, A. Kuchiki Ed., Institute of Developing Economies, Joint Research Program Series No. 141, Chiba, Japan. Keeble, D. and F. Wilkinson (1999) ‘Collective Learning and Knowledge Development in the Evolution of Regional Clusters of H-Tech SMES in Europe.’ Regional Studies, Vol. 9, No. 4, pp. 295–303. Kenney, M. (2000) Understanding Silicon Valley: The Anatomy of an Entrepreneurial Region. Stanford, CA: Stanford University Press. Kuchiki, A. (2004) Prioritization of Policies: A Prototype Model of a Flowchart Method. IDE Discussion Paper, No. 10.

202 Antonio José Junqueira Botelho et al. Kuchiki, A. (2007) ‘Clusters and Innovation: Beijing’s Hi-Technology Cluster and Guangzhou’s Automobile Cluster,’ in The Flowchart Approach to the Formation of Industrial Cluster: Focusing on the Endogeneous R&D and Innovation Mechanism, A. Kuchiki Ed., Institute of Developing Economies, Joint Research Program Series No. 141, Chiba, Japan. Kuchiki, A. and M. Tsuji (2005a) The Flowchart Approach to Industrial Cluster Policy. Unpublished. Chiba, Japan: IDE JETRO. Kuchiki, A. and M. Tsuji (eds) (2005b). Industrial Clusters in Asia: Analyses of Their Competition and Cooperation. Basingstoke: Palgrave Macmillan and Chiba, Japan: IDE JETRO. Leyten, J. (2004) Directions for Future Socio-Economic Research on ICTs. The Netherlands: TNO-STB. May, C. (2002) The Information Society: A Skeptical View. New York: Routledge. Navarro, L. (2003) Industrial Policies in the Economic Literature: Recent Theoretical Developments and Implications for EU Policy. Enterprise paper No. 12. European Commission, Directorate General, Brussels. OECD (1996) Network of Enterprises and Local Economic Development: Competing and Co-operation in Local Production Systems. Paris: OECD. Okada, A. (2005) ‘Bangalore’s Software Cluster,’ in Industrial Clusters in Asia: Analyses of Their Competition and Cooperation, A. Kuchiki and M. Tsuji Eds, Basingstoke: Palgrave Macmillan and Chiba, Japan: IDE JETRO, pp. 244–247. Picchieri, A. (2002) ‘Tesi sullo sviluppo locale.’ Studi Organizzativi Vol. 3, pp. 69–88. Pinto, M. M. (1999) Aspectos da dinâmica do sistema nacional de inovação – Uma investigação a partir da análise estratégica de um programa mobilizador de esforços para a Inovação. PhD Thesis. Department of Industrial Engineering, Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro. Piore, M. and C. Sabel (1984) The Second Industrial Divide. New York: Basic Books. Powell, W. W. (1996) ‘Interorganizational Collaboration in the Biotechnology Industry.’ Journal of Institutional and Theoretical Economics, Vol. 120, pp. 197–215. Powell, W. W., K. Koput, J. Bowie, and L. Smith-Doerr (2002) ‘The Spatial Clustering of Science and Capital: Accounting for Biotech Firm-Venture Capital Relationships.’ Regional Studies, Vol. 36, No. 3, pp. 291–360. Ramella, F. and C. Trigilla (eds) (2006) Reti Sociali e Innovazione: i sistemi localli dell’informatica. Rirenza: Fizenze University Press. Sapir, A. et al. (2003) An Agenda for a Growing Europe: Making the EU Economic System Deliver. Brussels: Economic Commission. Saxenian, A. (1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Cambridge, MA: Harvard University Press. —— (2006) The New Argonauts: Regional Advantage in a Global Economy. Cambridge, MA: Harvard University Press. SEBRAE (2004) Arranjos produtivos locais: perfil das concentrações das atividades econômicas no Estado do Rio de Janeiro: SEBRAE. SEBRAE-RJ (2005) Diagnóstico do Setor de TI no Estado do Rio de Janeiro. Rio de Janeiro: SEBRAE RJ, mimeo. SEBRAE RJ/Secretaria de Estado de Desenvolvimento Econômico do Rio de Janeiro (SEDE) Rede de Tecnologia do Rio de Janeiro(REDETEC)/Grupo de Produção Integrada/ Escola Politécnica e COPPE/ UFRJ (2007) Estudo sobre o Setor de Telecomunicações no Estado do Rio de Janeiro. Série estudos (Edição Preliminar), Rio de Janeiro, March, mimeo. Shavinina, L. (ed.) (2003) The Handbook on Innovation. Oxford: Pergamon.

Rio de Janeiro Software Cluster 203 SISTEMA FIRJAN (2007) Bases de uma estratégia para o fortalecimento da indústria de software. Rio de Janeiro: FIRJAN/GTM. Storper, M. (1997) The Regional World: Territorial Development in a Global Economy., London: The Guilford Press. Swann, P., M. Prevezer, and D. Stout (1998) The Dynamics of Industrial Clustering. Oxford: Oxford University Press. Trigilia, C. (2005) Sviluppo locale: um progetto per l’Italia. Bari: Laterza. Ueki, Y. (2007) ‘Industrial Development and Innovation System of the Sugar and Ethanol Sector in Brazil’, in The Flowchart Approach to the Formation of Industrial Cluster: Focusing on the Endogeneous R&D and Innovation Mechanism, A. Kuchiki Ed., Institute of Developing Economies, Joint Research Program Series No. 141, Chiba, Japan.

6 Innovation through Long-distance Conversations? Experience from Offshoring-based Software Clusters in Bangalore, India Aya Okada

6.1

Introduction

In recent years, clustering has become the focus of debates on regional economic development among scholars and policymakers. Recent literature, particularly theoretical development in economic geography and regional economics, has expanded and built on Marshall’s classic theory of agglomeration economies (Fujita et al. 1999; Rosenthal and Strange 2004). The literature on agglomeration and clustering usually postulates that clustered firms and industries gain localized external benefits, such as increasing returns to scale, pooling of skilled labour and knowledge spillovers. Other authors stress that firms and industries in clusters benefit from regional innovation systems (RIS) or local innovation systems (LIS): that is, regions develop the ability to innovate, as various institutions in clusters closely interact, network and learn together (Breschi and Lissoui 2001; Breschi and Malerba 2001, 2005; Cooke 2001, 2005). One assumption underlying the various strands of this recent literature is that firms are more likely to innovate and create new knowledge when clustered (Porter 1990, 2000). The reasoning is that spatial proximity between firms facilitates their creation and exchange of tacit knowledge, which has become more and more crucial as codified knowledge has become easily replicable and ubiquitous (Cumbers and MacKinnon 2006). On the other hand, as globalization proceeds, more and more global firms, particularly those based in developed countries, opt to outsource much of their work abroad, mostly to low-wage economies such as China and India. No longer does outsourcing apply only to low value-added labour-intensive manufacturing activities: now, highly value-added activities, in both manufacturing and services, are outsourced, including highly knowledge-intensive activities such as R&D in sectors like information 204

Offshoring-based Software Clusters in Bangalore 205

and communication technologies (ICT) and pharmaceuticals.1 As is widely known, many global technology firms outsource their work to India. Indeed, many global high-tech ICT firms have recently set up operations there. And great many firms in developed countries, most notably in the US, have started ‘offshoring’ some or all of their business processes and R&D from India.2 The great success of the Indian software industry, and its remarkable export growth as a key driver for the country’s economic development, have inspired many other developing and emerging countries. Brazil, Mexico, the Philippines, Vietnam, Pakistan and roughly 100 other countries are now making their efforts to develop export-oriented software clusters.3 This offshoring model, however, challenges the concept of clustering, particularly the successful experiences of knowledge-based clusters such as Silicon Valley (Saxenian 1994), on two grounds. First, contrary to the claim in the literature on clustering, offshoring involves little intra-cluster faceto-face interaction between transacting parties, which is considered an important motive for clustering and an important enabling factor for innovation, because clustered firms’ clients are largely located across distance. Second, again contrary to the claim, RIS/LIS may play only small roles as a forum to generate and exchange knowledge, because firms in clusters, though agglomerated, may establish their own external linkages to create and transfer knowledge, learning and innovation, and to expand their markets. Given this divergence from the theory on clustering, then, what does happen in the offshoring-based software clusters in India whose clients are predominantly foreign? How have they developed their innovative capabilities without the face-to-face interactions with their clients that are considered so important in promoting clustering among innovative firms? What local initiatives, if any, might be effective in inducing innovation, when new ideas mainly come from outside the cluster, and in fact mostly from abroad? This chapter examines these questions through a detailed case study of key firms in Bangalore, India’s largest software cluster, which has recently seen remarkable growth as a global hub for software development and IT services. The phenomenal growth of the Indian software industry, especially in Bangalore’s software cluster since the 1990s, and particularly since the turn of the century, is widely known and well documented (D’Costa 2004; Okada 2005, 2008; Parthasarathy 2000). This study draws on personal interviews in December 2004 and January 2005, with 30 software firms located in Bangalore, including both Indian leading software firms and subsidiaries of global IT firms. In this chapter I argue against the popular assertion regarding the links between innovation and clustering: I found that the main sources of knowledge transfer and innovation among key firms in Bangalore’s software cluster are their external linkages outside the cluster rather than face-to-face interactions between firms within the cluster. Moreover, the rich pool of

206

Aya Okada

skilled labour, made available through clustering has played an important role in facilitating learning, that is, diffusing the knowledge brought in by foreign networks. Thus, the study suggests a need to consider the role of external linkages in promoting innovation, particularly in export-led and offshoring-based clusters, an element overlooked in the so-called flowchart approach proposed by Kuchiki (2005) that stresses the importance of linkages between firms and local universities/research institutes within clusters. This chapter is organized as follows: Section 6.2 reviews factors that promote innovation in knowledge-intensive clusters; Section 6.3 describes the nature of offshoring and outsourcing practices in Indian software clusters, with a particular focus on Bangalore; Section 6.4 examines intra-firm and inter-firm channels and networks forged by software service firms; Section 6.5 summarizes the main findings and discusses some implications for developing offshoring-based knowledge-intensive clusters in developing countries.

6.2 Innovation and networks in knowledge-intensive clusters Various strands of the literature on clustering have pointed to the positive relationship between clustering and innovation. First, the conventional notion of regional clustering has focused on the role of local networks of specialized firms in generating external economies. Such local networks tend to achieve economies of scale as clustered firms cooperate in producing specialized products; they often stimulate learning and innovation through their close interactions, which lead to knowledge spillovers (Breschi and Malerba 2001, 2005). Agglomerations facilitate inter-firm learning as the economy increasingly relies on the transmission of complex uncodifiable information and tacit knowledge. A notable example of this type of clusters is the ‘old economy’ industrial clusters such as automobiles, where lead firms develop close inter-firm linkages with their suppliers, which facilitate inter-firm learning (Okada 2004; Okada and Siddharthan 2008). Secondly, other scholars argue that clustering induces more innovative activities by clustered firms, through the creation/presence of the so-called regional innovation systems (RIS) or local innovation systems (LIS), which involve collective learning among various local institutions such as firms, universities, training centres, R&D centres, science parks and government agencies, facilitating the generation and transfer of knowledge (Breschi and Malerba 2001, 2005; Cooke 2001, 2005). In a variant of this body of literature, Kuchiki (2005) proposes a ‘flowchart approach’ to the formation and development of industrial clusters. He suggests the importance of having adequately sequenced policy-induced conditions in place to foster clustering. This involves a series of public policies to attract lead firms, develop related industries, and upgrade human resources and

Offshoring-based Software Clusters in Bangalore 207

infrastructure. Another variant of this body of literature emphasizes the importance of informal personal networks among key local entrepreneurs, firm managers, engineers and workers, based on trust and shared norms and values; these networks facilitate learning and innovation through cooperation (Piore and Sabel 1984; Pyke et al. 1990). The existence of a professional community among key players who have often similar backgrounds helps develop such informal personal networks, as in the case of Silicon Valley (Saxenian 1994). Thirdly, still others focus on the competitive environment within a cluster as the major stimulus for the clustered firms to innovate (Porter 1998). Clustering may facilitate firms’ learning as they observe and monitor the activities of other firms within the cluster (Isaksen 2006). Clustered firms tend to gain access to specific information by being in a place where many other firms with related and complementary skills and knowledge operate; this may facilitate copying, learning and incremental innovation. The literature on knowledge-intensive clusters such as software and biotechnology emphasizes the importance of close contact and interactions, often face-to-face, between firms, clients and suppliers (Isaksen 2006; Storper and Venables 2005). Isaksen (2006), studying a software consultancy cluster in Oslo, found that the time-bound project nature of software consultancy enhances a need for face-to-face contact with various collaborators for learning and innovation, thus encouraging consultancy firms to be closer to their client firms. This is interesting, because clustering has been considered to occur among firms that seek economies of scale, by being closer to transacting firms such as suppliers and developing longterm trust-based linkages with them. Isaksen’s study, however, implies that regardless of whether contractual relations are short- or long-term, and regardless of the sectoral differences in organizing production processes and technologies, firms do commonly benefit from direct face-to-face interactions with their clients which stimulate knowledge transfer, learning and innovation. However, this claim draws on the case where most software firms have clients within the cluster. How relevant is this claim to cases like Bangalore, where most software firms have customers not within the cluster but in foreign markets on the other side of the globe? If close face-to-face interactions are so crucial for inducing innovation, how do offshoring-based software firms in emerging-economy clusters, such as those in Bangalore, overcome this difficulty? In this connection, some scholars emphasize external linkages to the global economy outside the clusters, particularly clusters in knowledgeintensive industries, as they bring in new ideas, knowledge and skills to the clusters (Saxenian and Hsu 2001). Particularly, in the context of emergingeconomy clusters, the role of foreign direct investment (FDI) may play an important role in facilitating knowledge transfer by MNCs. Indeed, knowledge workers and entrepreneurs are known as being ‘among the most

208 Aya Okada

widely connected or mobile people, always on the move and dependent on distantiated connections’ (Isaksen 2006, p. 190). Yet, these mobile knowledge workers tend to share similar norms, academic backgrounds and work experience, often thus creating the professional community, which in turn facilitates social interactions and provides a network of knowledge sharing. As Isaksen (2006) puts it, flows of goods, services, people, information, knowledge and technology occur within MNCs, between firms and project teams scattered around the world, across cluster boundaries and across national borders (p. 190). Given all this, can such external linkages help overcome the absence of face-to-face contacts for firms in offshoring-based knowledge-intensive clusters? And can they promote innovation? With respect to innovation, a recent study by Lester and Piore (2004) suggests that the ability to innovate means generating a stream of new products, improving upon old ones, and producing existing products in more efficient ways; doing this depends on the two fundamental processes of analysis and interpretation, so the key to sustaining innovativeness is finding a balance between the two processes (p. 6). They say that while innovation involves analytic process, mostly in the form of problem solving, it also involves open-ended interpretive processes, especially when the possible outcomes are unknown and ambiguous. Just as consumers do not always know what they want and need, engineers do not always know what the problems they have to solve (Lester and Piore 2004, p. 36). Hence, innovation necessarily involves some accidental, random, and experimental elements, and a process of interpretation, during which conversations among people and organizations with different backgrounds and perspectives help identify their problems and generate new ideas, which are both ambiguous at the outset (Lester and Piore 2004, p. 49). They identify the key differences between the analytical and interpretive approaches to product development, as summarized in Table 6.1. Thus, following this framework, for those engaged in offshoring, it is important to ensure that mechanisms that can facilitate this process of interpretation occur across distant locations and across national borders. Managers would need to facilitate conversations among people of different disciplines, technical backgrounds and cultures, across organizational and national borders. This suggests the importance of having some interfaces that can solicit interpretive conversations between firms, between IT professionals with different backgrounds and between engineers and their customers. In this respect, Lester and Piore (2004) suggest that a key issue in innovation, and in new product development in particular, is integration. They identify two kinds of integration. One occurs among technical specialists and across the boundaries of the different firms and organizational units involved in design and production; the second occurs along the boundaries between producers and final consumers (Lester and Piore 2004, p. 24).

Offshoring-based Software Clusters in Bangalore 209

Table 6.1 Analytic and interpretive perspectives Analysis

Interpretation

• The focus is a project, with a welldefined beginning and end • The thrust is to solve problems • Managers set goals • Managers convene meetings and negotiate to resolve different viewpoints and eliminate ambiguity • Communication is the precise exchange of chunks of information • Designers listen to the voice of customers • Means and ends are clearly distinguished and linked by a causal model

• The focus is a process, which is ongoing and open-ended • The thrust is to discover new meanings • Managers set directions • Managers invite conversations and translate to encourage different viewpoints and explore ambiguity • Communication is fluid, contextdependent, undetermined • Designers develop an instinct for what customers want • Means and ends cannot be clearly distinguished

Source: Lester and Piore (2004, p. 98).

Thus, they stress the important role of boundary management in corporate integration efforts. The above discussion suggests that we need to understand how firms manage their organizational boundaries, between different units within the organization and between firms; I focus on this process in the next sections.

6.3

Patterns of offshoring in Bangalore’s software cluster

This section briefly describes the patterns of offshoring commonly undertaken by software firms in Bangalore, India. Several scholars have identified salient factors contributing to the formation and development of Bangalore’s software cluster (Okada 2005; Parthasarathy 2000; Saxenian 2001). Among them are a large pool of skilled labour and very dynamic local labour markets; the historical agglomerations of high-technology firms and research institutions; historically cosmopolitan cultural environments and pleasant year-round climates; active government interventions at both national and state levels; and the successful experience of first-mover FDI firms, which have attracted other FDI firms (see Okada 2005 for details). Bangalore is receiving increasing attention as an outsourcing destination for global firms. In this section, I consider the nature of software development and services, and offshoring work in particular, focusing on activities that promote innovation.

210

Aya Okada

Containing over 1,300 software firms by 2003–2004, Bangalore’s software cluster has grown remarkably since 2000, largely driven by several large Indian firms, such as Tata Consultancy Services (TCS), Infosys Technologies, Wipro Technologies, Satyam Computer Services and HCL. These firms have transformed themselves to become global MNCs, with operations in many countries, by offering a wide range of contracted software services in outsourcing. Bangalore has also housed over 240 global technology MNCs, and over 600 Indian SMEs (see Okada 2005, 2008 for more detailed discussions on the structure of Bangalore’s software cluster). 6.3.1

The nature of software development and services

Before discussing the patterns of offshoring in Bangalore’s software cluster, it may be useful to review the nature of software development and services work. The software industry encompasses a wide range of activities that may be divided into five subcategories: (1) platform suppliers; (2) product development; (3) application software development; (4) software solutions and consultancy; and (5) after-sales services (support, maintenance and training).4 This section does not consider non-software activities such as IT-enabled services (ITES) and business process outsourcing (BPO), which include a wide range of back office activities using IT, although many software firms engage in ITES/BPO. Below, I describe each of the five types of activities. Though firms often engage in several of them simultaneously (Okada 2005), it is useful to understand them separately, as they often determine the organizational boundaries between firms, or between divisions within a firm. 6.3.1.1 Global technology firms and platform suppliers Global technology firms, particularly platform suppliers, deliver generic technology and tools that other firms use as the basis for developing software solutions (applications) by other firms (Isaksen 2006, p. 195). They are mainly large, global (and mainly US-based) IT firms, like IBM, Microsoft and Oracle, with subsidiaries and branch offices in various locations, including Bangalore. Most Indian firms, including many SMEs, in Bangalore use the platform technology to develop their own products and applications. For these global platform suppliers, the key to competitiveness is their large R&D efforts, which occur mainly in their US headquarters and increasingly in their offshore captive centres (subsidiaries) in India, most notably in Bangalore. Among software firms in the Bangalore cluster, these captive centres of the global technology firms such as Texas Instruments (TI) and Hewlett Packard (HP) carry out the most innovative software work – more innovative than giant Indian firms such as Infosys and Wipro. By developing technology and selling licenses, platform suppliers are able to engage in other activities such as consulting services, and customer support and maintenance.

Offshoring-based Software Clusters in Bangalore 211

6.3.1.2 Product development Developing software takes time and risk; it requires considerable initial capital, a large hardware market, and extensive distribution channels. Moreover, without internationally recognized brands, Indian software firms find it hard to export packaged software products in global markets. In addition, for their software products to be competitive, firms need to develop deep domain knowledge, coupled with strong managerial and marketing skills (Carmel and Tjia 2005). Therefore, few Indian software firms are engaged in developing proprietary or packaged standard software products. But many allocate a small proportion of their business to product development as an investment for future expansion of their businesses; meanwhile they focus on application software development and consultancy as their main source of revenue. Product development often takes the form of continuously upgrading existing products (Isaksen 2006, p. 196). Most large Indian software firms carry out offshoring-based R&D projects for foreign firms. For example, TCS has over 700 engineers in its R&D centre in Pune (in the state of Maharashtra); many work on such areas as software engineering, language processing, design, new product development and applied research.5 Likewise, Wipro has over 6,500 engineers in its software R&D division, which takes on R&D projects for foreign customers (Carmel and Tjia 2005). And roughly 200 engineers in Infosys’s R&D unit are grouped by their areas of activity, such as software engineering technologies, e-commerce, telecommunication and domain research.6 6.3.1.3 Application software development Application software development entails adapting and customizing standard software developed by other firms to meet the requirements of each customer firm. This includes consulting services, such as installation, integrating new solutions in the customer firms, training employees of the customer firm, converting existing data to a new software system (Isaksen 2006, p. 196). This is the largest category of activities carried out by Indian firms, both large ones and SMEs. Software development involves several phases, including identifying and analysing client needs, prioritizing among various demands, deciding on specification of software, designing the software, programming, testing and delivering. Of these phases, Indian firms initially focused on low-end application software development, particularly in low-end design, programming and coding. 6.3.1.4 Software solutions and consultancy Consultancy projects entail the development of customized solutions, such as developing and implementing a new software system for a client firm (Isaksen 2006, p. 196). Top Indian software firms, such as TCS, Infosys and Wipro, have all achieved extraordinary successes in offshoring-based

212

Aya Okada

software solutions and consultancy services; by 2003, they had captured business with more than half of the US Fortune 500 firms (NASSCOM 2004). Consultancy projects often include advice on the purchasing and installing software products; analysing businesses’ processes, competence and business needs; organizing business activities for client firms, and preparing IT strategies (Isaksen 2006, p. 196). These firms often develop new customized software solutions, building on those developed in previous projects. Firms may develop standardized programmes or solutions based on successful projects, involving the process of converting tacit knowledge into codified knowledge. They serve client firms from a wide range of sectors including insurance, finance, banking, telecommunications, defence and the public services, and increasingly many manufacturing firms. Firms that have long-term projects with certain clients send their workers to be placed on site at those firms. Consulting firms develop their competitiveness in several ways: they continuously build on their competence and skills, their workers continuously update their knowledge, and they seek exposure to the latest technologies in the field (Isaksen 2006, p. 197). Firms often assemble teams that include the different skills sets required to complete the tasks, for each project (Okada 2005). They often hire new workers with specialized skills if they cannot find the appropriate skills internally. Some members of the project may spend time on site at a client firm, especially at the early stage, to understand the client’s requirements and ways of conducting business. At the same time, these consulting firms often provide extensive internal training to their employees to update their skills and knowledge. New knowledge is diffused internally within the firm through such training and websites. 6.3.1.5 After-sales services (support, maintenance and training) Most software firms carry out after-sales services: they provide customer support, maintain software systems and train the client’s employees to use the systems. While these after-sales services often entail fairly standardized procedures, they also serve as a feedback mechanism to provide new information from customer firms back to their product development work. Typically, Indian software services firms, such as Infosys and TCS, organize their operations and project units in a matrix; some types of activities (product development, application development, solutions and consultancy, after-sales support) are on one axis, and various industrial domains (banking, financial services, insurance, telecommunication, manufacturing) are on the other.7 Project teams are assembled based on this combination of skill sets, which also determine the organizational boundaries between units and between projects. With respect to contracts, most software development falls into one of two models: either a short-term project, often on a one-by-one basis, or an ongoing contract with particular clients. One form of the latter is an

Offshoring-based Software Clusters in Bangalore 213

offshore development centre (ODC). A software service provider firm dedicates a software development centre to a foreign client firm; the client may offer some specialized hardware and software, while the provider may allocate dedicated staff and other resources just for this client (Carmel and Tjia 2005, p. 105). 6.3.2 Patterns of offshoring-based software development and services Several factors explain the recent phenomenal growth of offshoring practices widely observed in many industries and services, especially software. First, the recent remarkable technological advances in ICT have dramatically reduced communication costs almost to zero. Secondly, faced with intensified global competition, firms in developed countries, most notably those in the US, have had to respond to growing cost pressures, making offshoring a critical strategic necessity (Carmel and Tjia 2005). Thirdly, on a related point, low-wage economies like India and China have large pools of English-fluent and well-trained yet low-wage scientists and engineers. Fourthly, time differences between the US and India make it possible for US-based global technology firms to work around the clock. Finally, in recent years, software development practices and tools have become so standardized that software tools have become nearly undifferentiated by producer (Carmel and Tjia 2005, p. 4). Given the high level of standardization, can offshoring practices allow Bangalore’s software cluster to develop innovative capabilities? Indeed, in an earlier nascent phase in the 1990s, many Indian software firms were mainly engaged in low-end labour-intensive tasks such as data conversion, software customization, website development and hosting, and reuse of codes, all of which have become increasingly standardized. Activities such as bug-fixing (in the maintenance phase) and after-sales technical support are also considered suitable for offshoring, because they are small tasks, of low complexity, and can be routinized between different sites (Carmel and Tjia 2005, p. 12). Indian firms initially entered this segment in the global (mainly US) market, by sending their lower-wage Indian engineers to their customers’ premises on site, in a practice dubbed ‘bodyshopping’ (Parthasarathy 2000). As Bangalore’s software cluster grew, however, and its software firms proliferated, the practice of offshoring also expanded, not only among subsidiaries of MNCs, including platform suppliers and large Indian software firms, but also among many Indian SMEs. Foreign clients of these offshore Indian firms are not only software end-users (such as banks, airlines and manufacturing firms) but software firms, which in turn provide services to end-users. In the latter case, Indian firms subcontract, taking parts of a large project on a small project basis.8 Table 6.2 shows the changing patterns of modes of delivery of software services exported from India.

Table 6.2

Modes of delivery of software services exports from India (%) Year 1994–1995

1998–1999

1999–2000

2000–2001

61.0 29.5 9.5

58.18 33.92 7.90

57.43 34.70 7.87

56.08 38.62 5.29

On site Offshore Products and unclassified Total Source: Sridharan (2004).

100

100

100

100

2001–2002 45.21 50.68 4.11 100

2002–2003 38.95 57.89 3.16 100

Offshoring-based Software Clusters in Bangalore 215

Software work is commonly divided into about ten stages: (1) conceptualization; (2) requirement analysis and specification; (3) high-level (integrated) design; (4) low-level (detailed) design; (5) coding and programming; (6) prototyping; (7) unit testing; (8) delivery and integration; (9) system testing; and (10) customer support and maintenance (see Figure 6.1). As Figure 6.1 shows, the work flow is actually a reversal of the value chain flow: The more upstream activities a firm is engaged in, the more it moves up the value chain. Among these stages, those most often offshored are: (4) low-level (detailed) design; (5) coding and programming; (6) prototyping; (7) unit testing; and (10) customer support and maintenance. These activities are mostly standardized, can be precisely defined and specified, may be considered tedious, repetitive and undesirable (Carmel and Tjia 2005, p. 14). Indeed, many offshore activities are single projects that are contracted on a piecemeal basis, rather than the entire processes being transferred to outsider firms (Carmel and Tjia 2005). According to a small Indian software firm in Bangalore, activities such as low-end (detailed) design, coding and programming follow the standard guidelines and procedures set by the industry such as SEI CMM.9 These activities thus largely rely on codified knowledge; therefore, according to Lester and Piore (2004), they are analytic (see Table 6.1 in Section 6.2). By contrast, activities that tend to stay onshore, often in the headquarters of MNCs, are: (1) conceptualization (i.e., what client firms want); (2) analysis of customer requirements (i.e., what product will be needed, and how products should look like); (3) high-level (integrated) design; (8) delivery and integration; and (9) system testing. These activities require little standardization, but they do require close, and often face-to-face, interactions with customers. They also require solid domain knowledge, and deep cultural knowledge, because of the need to meet with clients and Onshore ① ② ③ ④ ⑤ ⑥ ⑦ ⑧ ⑨ ⑩

Conceptualization Requirements analysis & specification High-level (integrated) design Low-level (detailed) design Coding/programming Prototyping Unit testing Delivery/integration System testing Customer support/maintenance

Offshore

Work Value chain flow stream

x x x x x x x x x x

Figure 6.1 Software development services life cycle by location of activities Source: Personal interviews with Infosys (December 2004), adapted from Carmel and Tjia (2005).

216

Aya Okada

talk to them in their own language (Carmel and Tjia 2005, pp. 14–15). These processes require extensive conversations between engineers who have technical knowledge and people who have solid domain knowledge, and between software development firms and their customers. Thus, according to Lester and Piore (2004), they involve interpretive processes (see Table 6.1 in Section 6.2). As Figure 6.1 shows, these are also higher-end activities that add higher value, as they are more creative, innovative and research-oriented, requiring broad knowledge (Carmel and Tjia 2005, p. 15). These activities constitute an important part of the consulting work discussed in the previous subsection. Global technology firms based in developed countries tend to keep these activities at home (onshore) to maintain their competitiveness (Carmel and Tjia 2005, p. 15). But some Indian software firms, and even SMEs, are taking on product development projects, however small, going through the entire process from conceptualization to final product release. And many Indian firms, including SMEs, want to expand the scope of their activities by moving up to higher-end activities. The offshoring model poses some important challenges to both the clients in developed countries and their service provider firms in India. Carmel and Tjia (2005) identify five important challenges: (1) It is difficult to communicate over distance; to convey some ambiguous ideas and tacit knowledge using e-mail or the telephone. (2) Coordination can be difficult; software development requires a series of adjustments between the firm and its customers. People working on a software project normally coordinate with their team members on the project and their customers through numerous adjustments, which often occur through ‘spontaneous, face-to-face conversation’ (Carmel and Tjia 2005, p. 12). Obviously, offshoring makes such conversations difficult, so necessary adjustments are harder. (3) Monitoring and controlling software production processes can be hard: ‘successful management control takes place when managers can roam around to see, observe, and dialogue with their staff’ (Carmel and Tjia 2005, p. 12). But, with offshoring, managers can hardly supervise their staff this way. (4) It is hard to develop social bonds; offshoring makes it difficult to create a sense of teamwork and build trust among project members who are so far away from each other. (5) Cultural differences interfere; offshoring inevitably places the staff in cross-cultural settings, requiring them to develop cross-cultural communication and understanding, which is hard to come by. Given that the division of labour within the global software industry takes place along these phases between onshore and offshore locations, and given these difficulties in offshoring operations, Indian software clusters have little scope to become innovative, unless they engage in higher-end activities and move up the value chain. Indeed, given that India’s software industry has concentrated heavily on relatively low-end services, some authors even argue that the Indian software industry is not innovative, as it

Offshoring-based Software Clusters in Bangalore 217

is constrained by a very small domestic hardware market (D’Costa 2004). If this is true, have software clusters in India, especially in Bangalore, grown without innovation, totally challenging the notion of close links between clustering and innovation? But if Indian software firms provide offshoring services involving little innovation, it raises a further puzzle: why have so many global technology firms agglomerated in Bangalore and why many Indian software firms have grown so remarkably to become software service firms offering contracted R&D activities to the world’s top companies? Indeed, global technology firms, mainly platform suppliers, which engage in higher-end activities, go beyond cost reduction concerns and benefit from innovation, speed and flexibility, through offshoring from India (Carmel and Tjia 2005). These firms engage in offshoring to increase their speed and flexibility, taking less time to complete projects, and responding quickly to rapidly changing business needs, and improving innovation capabilities by tapping highly skilled and creative engineers. Indeed, by 2003, 77 global software product firms had established direct R&D subsidiaries in India (Carmel and Tjia 2005, p. 11). A question remains: how have software firms acquired and developed their innovative capabilities? The next section addresses this question.

6.4 External linkages of software firms in Bangalore This section addresses the channels and mechanisms through which knowledge for innovation is transmitted to software firms in Bangalore, with particular focus on their external linkages. The recent literature on clustering and networks points out the important role of external linkages in clustering and innovation (Breschi and Malerba 2001, 2005; Cooke 2001, 2005; Saxenian and Hsu 2001). For emerging clusters such as Bangalore, external linkages may allow clustered firms access to knowledge, skills, contacts, capital, information on customer demands, technologies and market trends. On the other hand, for established and mature clusters such as Silicon Valley, external linkages with other regional clusters allow clustered firms to upgrade the industrial base and reduce the risk of lock-in by keeping the cluster open to new ideas and technologies (Breschi and Malerba 2005). In other words, such external linkages may facilitate ‘the conversation’ among entrepreneurs and knowledge workers across different geographical locations (Lester and Piore 2004), creating a forum for exchanging ideas and interpretations, which thus may induce innovations. Bangalore firms are engaged in five types of external linkages: (1) intrafirm linkages between subsidiaries of MNCs (including platform suppliers); (2) inter-firm linkages, including alliances with global firms; (3) international movement of highly skilled technical and managerial workers; (4) deployment by software firms of their staff to client firms abroad on a

218 Aya Okada

fixed-term basis, ranging from a few weeks to a few years; and (5) the extensive use of agents, particularly for Indian firms, who serve as interfaces and interpreters, and transmit knowledge to software firms in Bangalore. In the rest of this section, I discuss the experiences of firms in Bangalore, with respect to each of these types. 6.4.1 Intra-firm linkages and global division of labour within MNCs Both Indian-based and US (or European)-based MNCs operating in Bangalore have their own global networks. For example, TCS, the largest Indian software firm, which was established in 1968 and had a total revenue of US$1.56 million in 2003–2004, has operations in 47 countries, with 152 offices across the globe.10 Within India, it also has operations in seven clusters beside Bangalore: Delhi, Mumbai, Pune, Chennai, Hyderabad, Calcutta and Lucknow. Each location specializes in different types of activities, exhibiting interesting patterns of intra-firm division of labour. All its locations within India are connected through its Intranet. TCS provides mostly foreign clients with diverse IT consulting and services in the areas of IT technology and infrastructure, architecture, design and development; testing and deployment; and systems integration; and application management. In particular, by 2005, TCS had set up 33 clientdedicated offshore development centres (ODCs), spread over ten cities in North America, Budapest, Melbourne, Montevideo, Guildford (England), Haungzhou and Yokohama.11 Similarly, Infosys and Wipro have dedicated R&D centres for particular foreign clients. These ODCs and R&D centres dedicated to foreign clients clearly help facilitate close interactions between these Indian software firms and their clients including leading global technology firms. Thus they facilitate the interpretive processes described earlier, and in turn they help them develop innovative capabilities. A large pool of highly skilled IT professionals available in Bangalore (see Okada 2005 for detailed discussions) has enabled these global firms to develop a well-coordinated intra-firm global division of labour, efficiently deploying the best talents to specific types of projects even in niche areas. These talented software engineers are key to competitiveness and innovation in software R&D. For example, 7,000 to 9,000 engineers are involved in IC chip design in India; about 60 per cent of them work in the Bangalore cluster.12 The large supply of highly skilled software engineers has clearly attracted many MNCs to locate in Bangalore, including their R&D centres. This has led to considerable knowledge and skill transfers to local firms in the cluster (Patibandla and Peterson 2002). In turn the agglomeration of these leading global IT firms has led many well-trained software engineers to gather in Bangalore (Okada 2005), creating a virtuous circle. Among software firms in the Bangalore cluster, these captive centres of global technology firms carry out most innovative software work – more innovative than

Offshoring-based Software Clusters in Bangalore 219

activities done by most successful Indian software services firms such as Infosys and Wipro. For example, Texas Instruments (TI), the first MNC to set up an offshore operation in Bangalore back in 1985, takes advantage of elaborate intra-firm division of labour linking its many operations across different countries. Figure 6.2 illustrates the pattern of its intra-firm division of labour across countries, which the firm introduced to take advantage of the ‘24x7’ design cycle, and to reduce production costs. Within TI, however, the most innovative function – silicon research – is still carried out only in the US.13 Like TI, most global technology firms have their own arrangements of intrafirm division of labour involving their overseas operations across countries, making use of the locational advantages of each operation. It is worth mentioning here that global platform suppliers and technology firms also do forge internal linkages, within the cluster, with those local software firms that are users of their platforms. For example, Microsoft has its own certification schemes to select local software firms that depend on Microsoft platforms, and organizes seminars for these local users to diffuse knowledge related to its product.14 Likewise, Intel has developed partnerships with its product users through the firm’s ‘Early Access Program’, which allows local users to test its new product before it goes on the market so that it can get some instant feedback from them. Intel responds to comments Activities

Locations

Specification/design

USA

IC chip design/development

India

Production

China/Taiwan

Packaging

Korea

Packing

Final products

Malaysia

To Europe

Figure 6.2 Pattern of intra-firm international division of labour: The case of TI’s semiconductor production Source: Personal interviews with a senior manager of TI (January, 2005).

220

Aya Okada

received from these local users by rectifying the problems they identify. In this way, knowledge transfer and diffusion also occur between these global technology firms and local software firms within the cluster. Also, these global platform suppliers and technology firms use many local firms, mainly SMEs, as subcontractors for low-end software development activities such as low-end designs and programming. These tasks are highly standardized and routinized, and thus leave little room for local subcontractors to engage in innovation activities. In fact, Intel’s software development is almost entirely carried out by local Indian firms, including large giants such as Wipro and Infosys, while Intel maintains its chip design work – higher-end activities – internally. Similarly, about 70 per cent of Microsoft’s work is subcontracted to local software firms on a project basis.15 In other words, global platform suppliers located in Bangalore forge vertical inter-firm linkages within the cluster. But, these vertical linkages differ from those in ‘old economy’ manufacturing clusters, where large MNCs firms play a leading role in developing their supporting industries within the cluster by forging backward linkages with local firms. Interestingly, by contrast, in Bangalore’s software cluster, large global MNCs create both forward and backward linkages with local Indian firms. Internal vertical backward linkages created within the cluster between MNCs and local firms, however, do not lead the latter to develop their innovative capabilities, because the work subcontracted to local firms is so standardized, repetitive and tedious. 6.4.2 Inter-firm external linkages and alliances Many firms in Bangalore forge extensive external links with firms and other institutions outside the cluster. For example, IBM, a leading global platform supplier, entered into a strategic alliance with i-flex, an Indian software product development firm, producing a banking software product on the IBM platform. IBM provides technical support for i-flex, as it does for many other software firms that use the IBM platform. These arrangements facilitate knowledge transfer from the global platform supplier to the Indian firms. Many Indian firms, including SMEs, use another strategy to move into higher-end activities and to become more innovative: merger and acquisition (M&A). For example, Eximsoft Ltd, a Bangalore-based Indian mediumsized firm providing relatively low-end offshore software services to foreign clients, entered into a M&A deal and was acquired in December 2004 by a US-based technology consulting firm, to grow into a high-end consulting and technology services firm with greater geographic coverage for its market.16 This kind of alliance allows Indian SMEs to gain access to higher levels of domain knowledge, managerial skills and markets. Moreover, it facilitates ‘conversations’ between professionals who have different

Offshoring-based Software Clusters in Bangalore 221

Table 6.3

TCS’s partnerships with universities worldwide

Universities in partnership

Joint R&D projects

University of Wisconsin, Milwaukee University of California, Los Angeles University of California, Riverside University of Humberside, UK Rotterdam School of Management IIT Chennai, Mumbai, Kharagpur Indian Institute of Science, Bangalore Carnegie Mellon University SIMTech., Singapore Aalborg University, Denmark Simon Frazer University, Vancouver State University of New York King’s College London, UK

Database technologies Multimedia Wireless technologies Systems engineering & Cybemetics e-Business VLSI Design, Intelligent Internet APDAP centre Emerging trends in the software industry Software products in embedded systems 4G mobile technology IT research & development Supporting real time profiles

Source: TCS’s unpublished internal document (2005).

backgrounds, business cultures and experiences; this generates more creative ideas and inspiration and thus is conducive to innovation. In addition, some Indian firms have forged partnerships with foreign universities and research institutions. For example, TCS has established partnerships with over 12 universities around the world, including the University of California at Los Angeles, the University of Wisconsin, Milwaukee, and Carnegie Mellon University, for collaborative technology development (see Table 6.3). In addition, TCS annually sends five to ten employees to these universities for study and joint research.17 Clearly, these partnerships provide TCS with access to the latest technologies and allow the firm to engage in innovation. Interestingly, while the literature emphasizes the important role of firm–university links within the cluster as part of the local innovation systems (LIS) as discussed in Section 6.2, the experience of TCS shows that its links with universities outside the cluster across distance actually help the firm develop its innovative capabilities. 6.4.3 International mobility of a highly professional workforce Bangalore’s software cluster benefits from a high level of international mobility, as highly skilled IT professionals move between India and other nations. This mobility arises from two sources: First, as is well known, Indian engineers who have studied and worked in foreign countries, especially the US, return to India. Secondly, Indian professionals are assigned to foreign client firms to carry out on-site work, on a short-term or longterm basis, often as a part of their projects, as discussed in the next subsection.

222

Aya Okada

Returnees from the US do not yet constitute the kind of large professional community in Bangalore (Okada 2005) that they do in Taiwan (Saxenian and Hsu 2001). Indeed, even TCS, the largest Indian software firm, which has a staff of over 40,000 associates/engineers, has virtually no PhD holders. About 60 per cent of its engineers hold a Bachelor’s degree, and 36 per cent hold Master’s and MBA degrees; these are mostly graduates from India’s best educational institutions. However, among the young entrepreneurs who are starting India’s software SMEs, it is increasingly common to find some returnees who were educated in a foreign (mainly US) university and have worked in large technology firms in the US before they returned to India. For example, a small Indian software services firm was established in 2000 by three young entrepreneurs who are friends, including one engineer who worked with Microsoft at its Richmond headquarters for eight years.18 Moreover, even at small Indian software development firms a considerable number of employees have foreign experiences. For example, of the 70 employees at a medium-sized firm established in 2003, 12 have overseas work experience, and the owner noted that their experiences are valuable to the firm.19 A more common practice is intra-firm transfer of Indian professionals of MNCs, Indian large firms and even medium-sized firms, between India-based operations and their overseas operations. For example, even a medium-sized Indian software firm in Bangalore, one with 150 employees, assigns 40 of them to its office in Japan.20 6.4.4 The role of external agents Also, many firms claim that demands from their customers, often in the US market, are an important source of innovation.21 This means that not only supply conditions but also demand factors are important in influencing software firms’ innovation activities. Clients of exporting firms are located across national borders, often on the other side of the globe, which obviously makes face-to-face interactions difficult. To overcome this problem, most Indian software exporting firms use some agents, as intermediaries, in their foreign markets, such as the US, Europe, Japan and sometimes in Singapore, Malaysia and China. These agents act as an interface between the Indian software firms and their overseas clients. These agents may be firms’ branches, sales offices and subsidiaries; in other cases they are personal friends, relatives or a few marketing staff. Virtually all the 30 firms that I interviewed had some of these overseas agents, often in multiple locations. They are often Indian, and sometimes the non-resident Indian (NRI). For example, a small Indian software firm with 40 employees, established in 1996, has an associate in Singapore and another in the UK. These associates understand the needs of clients, monitor the projects at clients’ sites, and also look for new clients.22 Another small Indian firm has 11 associates abroad, mostly part-time independent consultants, handling 10 to 12 clients each;

Offshoring-based Software Clusters in Bangalore 223

they provide close monitoring and customer support, and identify new clients.23 Another firm, established in 1999, has 35 employees; it has two agents in Malaysia and the US, who are the owner’s relatives, and also collaborates with independent associates on a contractual basis. The owner says these overseas networks help transfer technical knowledge to the firm, provide it with financial support and help identify new clients and source materials.24 These agents visit customers on site to discuss and understand the clients’ problems, which they then convey to the Indian firms. They understand not only the nature of their clients’ business but also the larger business environments, market trends and clients’ languages and culture. Thus they are effective as an interface in facilitating cross-cultural and cross-sectoral ‘conversations’ between the Indian firms and their clients across distance. In turn they facilitate interpretive processes, because these agents convey not only technological information and domain knowledge, but also demands and feedback from customers.

6.5 Conclusion This chapter, drawing on a case study of Bangalore’s software cluster, has examined how firms in offshoring-based knowledge-intensive clusters in India have developed their innovative capabilities, despite having little face-to-face interaction with their predominantly foreign clients, a factor considered to promote clustering among innovative firms and to facilitate innovation. The study has first identified what types of software development and services activities are outsourced to Indian software firms in Bangalore by foreign client firms in mainly developed countries. It revealed that those offshored software development and services activities are mostly relatively low-end, labour-intensive, standardized and repetitive activities such as low-end (detailed) design, programming, coding and maintenance. Yet, in recent years, Indian software firms in Bangalore have diversified their types of activities in order to move up the value chain; more firms are now increasingly engaged in higher-end activities such as product development and R&D, and even take on the whole life cycle of software development activities. This study found that leading software firms in Bangalore have devised various channels to engage in long-distance conversations with other divisions of the same firms, client firms, universities and agents. Together these channels facilitate the inflow of ideas, information and knowledge for promoting innovation, thus letting the firms move into more higher-end activities. Clustered firms in Bangalore have independently established five types of external channels with various individuals, firms and universities; these channels facilitate interpretive processes, and thus innovation. These are: (1) internal firm structures of MNCs (including the top Indian software services firms that themselves became MNCs) functioning across distance

224 Aya Okada

and across nations, and especially intra-firm linkages and global division of labour within MNCs across distance; (2) inter-firm external linkages and alliances, including links with top global technology firms in the forms of dedicated ODCs and R&D centres, alliances through M&A with foreign firms that carry out higher-end activities, and partnerships with foreign leading universities around the world for collaborative research and innovation; (3) internationally mobile highly professional workforce; Indian engineers and professionals who were educated abroad and have work experience with global IT firms abroad return to India, and Indian staff transfer to Indian firms’ overseas operations; (4) organizational arrangements to combine offshoring and on-site work as a boundary management strategy; and (5) the extensive use of agents placed in foreign markets; these may be individuals (often friends and relatives), independent organizations, and their branch offices, subsidiaries and sales offices. Acting as an interface between Indian software firms located in Bangalore and their foreign clients abroad, they facilitate ‘conversations’ between them, transferring and diffusing new ideas, knowledge and technologies, and thus helping the Indian firms in Bangalore improve their products, solutions and processes. The existence of external linkages enabled technological knowledge from demanding clients to be transmitted, and market opportunities to be expanded. On the other hand, global platform suppliers and multinational technology firms having operations in Bangalore forge both backward and forward linkages with local Indian firms, and SMEs in particular, both as end-users of their platform technologies and as subcontractors. However, the backward linkages through subcontracting involves mainly low-end, labourintensive and tedious activities such as programming and coding, leaving little room for these firms to develop innovative capabilities. Therefore, contrary to the popular belief that important links exist between innovation and clustering, this study finds that external international linkages and the internal firm structures are more important in inducing innovative activities for local software clustered in Bangalore. Therefore, it challenges the notion of localized knowledge spillovers through close inter-firm linkages within the cluster, which is widely cited in the literature as a critical factor that promotes innovation. The greater importance of the external linkages than internal face-to-face linkages within clusters may be explained by the very nature of Bangalore’s cluster as export-led and offshoring-based; demands for new knowledge creation and new product/ process development are generated across distance in foreign markets rather than inside the Bangalore cluster or in India. As discussed elsewhere (Okada 2005), the development of Bangalore’s software cluster is explained at least in part by the unique and path- dependent way that it evolved. Several other factors have also been crucial: a large pool of highly skilled labour and very dynamic local labour markets within the cluster, active government interventions, the historical development of

Offshoring-based Software Clusters in Bangalore 225

engineering and technology industries in Bangalore and a cosmopolitan atmosphere with a large English-speaking population and relatively pleasant climate throughout the year. However, for promoting innovation, factors such as the existence of external linkages need to be considered. The findings of this study thus suggest that while the literature on clustering tends to link clustering and innovation, and focuses on the local factors within the cluster, such as the presence of local innovation systems (LIS) as important determinants, factors that explain the formation and development of clustering may differ from those that explain innovation. This study therefore offers some implications for policy: For promoting innovation in developing-country clusters, particularly in those that are offshoring-based, governments may play an important role in facilitating the creation and expansion of external linkages between clustered firms and foreign actors such as firms, agents and universities, in addition to supporting the establishment of LIS.

Notes 1. Outsourcing means contracting tasks and processes to be performed outside the boundaries of the firm (Carmel and Tjia 2005). 2. Offshoring means the shifting of tasks to developing countries and emerging economies (Carmel and Tjia 2005). In contrast, on-site services mean work provided locally (on site) by foreign (developing-country) firms, often using lowerwage foreign (developing-country) workers. 3. Carmel and Tjia (2005) classify over 100 software exporting nations into three tiers, according to their stage in exporting software. Tier 1: mature softwareexporting nations, including developed countries, Israel, Ireland, India, China and Russia. Tier 2: emerging software-exporting nations, including Brazil, Mexico, Costa Rica, the Philippines, Malaysia, Sri Lanka, Pakistan and many Eastern European countries. Tier 3: infant-stage software exporting nations, including El Salvador, Jordan, Egypt, Bangladesh, Indonesia, Vietnam and many others. 4. This classification draws on Isaksen’s (2006) study, which classifies software work into four categories: (1) platform suppliers; (2) software production; (3) consultancy; and (4) after-sales services. 5. Personal interviews with TCS, Bangalore, December 2004. 6. Personal interviews with Infosys, Bangalore, December 2004. 7. Personal interviews with Infosys and TCS, Bangalore, December 2004. 8. Many firms I interviewed stated that they have this contracting pattern (e.g., Bright Sword Technologies, Bangalore, November 2004). 9. Personal interviews with Metalearn, Bangalore, November 2004. 10. Personal interviews with TCS, Bangalore, December 2004. 11. Personal interviews with TCS, Bangalore, December 2004. 12. Personal interviews with TI, Bangalore, January 2005. 13. Personal interviews with TI, Bangalore, January 2005. 14. Personal interviews with a small Indian software firm (Company M), Bangalore, November 2004. 15. Personal interviews with a small Indian software firm (Company B), Bangalore, November 2004.

226

Aya Okada

16. Personal interviews with several managers from Eximsoft and Trianz in Bangalore, December 2004. 17. Personal interviews with TCS, Bangalore, December 2004. 18. Personal interviews with a small Indian firm (Company B), Bangalore, November 2004. 19. Personal interviews with a medium-sized firm (Company I), Bangalore, December 2004. 20. Personal interviews with a medium-sized Indian firm (Company F), Bangalore, December 2004. 21. For example, personal interviews with TI, Bangalore, December 2004. 22. Personal interviews with PLSPL, Bangalore, December 2004. 23. Personal interviews with a small Indian firm (Company S), Bangalore, December 2004. 24. Personal interviews with a small Indian firm (Company L), Bangalore, December 2004.

References Breschi, Stefano and Franco Malerba (2001) ‘The Geography of Innovation and Economic Clustering: Some Introductory Notes.’ Industrial and Corporate Change, Vol. 10, No. 4, pp. 817–833. —— (2005) ‘Clusters, Networks, and Innovation: Research Results and New Directions,’ in Clusters, Networks, and Innovation, Stefano Breschi and Franco Malerba Eds, Oxford: Oxford University Press, pp. 1–26. Breschi, Stefano and Francesco Lissoni (2001) ‘Knowledge Spillovers and Local Innovation Systems: A Critical Survey.’ Industrial and Corporate Change, Vol. 10, No.4, pp. 975–1005. Carmel, Erran and Paul Tjia (2005) Offshoring Information Technology: Sourcing and Outsourcing to a Global Workforce. Cambridge: Cambridge University Press. Cooke, Philip (2001) ‘Regional Innovation Systems, Clusters, and the Knowledge Economy.’ Industrial and Corporate Change, Vol. 10, No. 4, pp. 945–974. —— (2005) ‘Regional Knowledge Capabilities and Open Innovation: Regional Innovation Systems and Clusters in the Asymetric Knowledge Economy,’ in Clusters, Networks, and Innovation, Stefano Breschi and Franco Malerba Eds, Oxford: Oxford University Press, pp. 80–109. Cumbers, Andy and Danny MacKinnon (2006) ‘Introduction: Clusters in Urban and Regional Development,’ in Clusters in Urban and Regional Development, Andy Cumbers and Danny MacKinnon Eds, New York: Routledge, pp. i–xvii. D’Costa, Anthony P. (2004) ‘The Indian Software Industry in the Global Division of Labour,’ in India in the Global Software Industry: Innovation, Firm Strategies and Development, Anthony P. D’Costa and E. Sridharan Eds, Delhi: Macmillan India, pp. 1–26. Fujita, Masahisa, Paul Krugman, and A. J. Venables (1999) The Spatial Economy: Cities, Regions, and International Trade. Cambridge, MA: MIT Press. Isaksen, Arne (2006) ‘Knowledge-based Clusters and Urban Location: The Clustering of Software Consultancy in Oslo,’ in Clusters in Urban and Regional Development, Andy Cumbers and Danny MacKinnon Eds, New York: Routledge, pp. 187–204. Kuchiki, Akifumi (2005) ‘A Flowchart Approach,’ in Industrial Clusters in Asia: Analyses of Their Competition and Cooperation, Akifumi Kuchiki and Masatsugu Tsuji Eds, New York: Palgrave Macmillan, pp. 169–199.

Offshoring-based Software Clusters in Bangalore 227 Lester, Richard K. and Michael J. Piore (2004) Innovation: The Missing Dimension. Cambridge, MA: Harvard University Press. NASSCOM (2004) Strategic Review 2003: The IT Industry in India. New Delhi: NASSCOM. Okada, Aya (2004) ‘Skills Development and Interfirm Learning Linkages under Globalization: Lessons from the Indian Automobile Industry.’ World Development, Vol. 32, No. 7, pp. 1265–1288. —— (2005) ‘Bangalore’s Software Cluster,’ in Industrial Clusters in Asia: Analyses of Their Competition and Cooperation, Akifumi Kuchiki and Masatsugu Tsuji Eds, New York: Palgrave Macmillan, pp. 244–277. —— (2008) ‘Small Firms in the Indian Software Clusters: Building Global Competitiveness,’ in High-tech Industries, Employment and Global Competitiveness, S. R. Hashim and N. S. Siddharthan Eds, London and Delhi: Routledge, pp. 43–69. Okada, A. and N. S. Siddharthan (2008) ‘Automobile Clusters in India: Evidence from Chennai and the National Capital Region,’ in The Flowchart Approach to Industrial Cluster Policy, Akifumi Kuchiki and Masatsugu Tsuji Eds, New York: Plagrave Macmillan, pp. 109–144. Parthasarathy, Balaji (2000) Globalization and Agglomeration in Newly Industrializing Countries: The State and the Information Technology Industry in Bangalore, India. PhD dissertation, Berkeley, CA: University of California. Patibandla, Murali and Bent Peterson (2002) ‘Role of Transnational Corporations in the Evolution of a High-Tech Industry: The Case of India’s Software Industry.’ World Development, Vol. 30, No. 9, pp. 1561–1577. Piore, Michael J. and Charles F. Sabel (1984) The Second Industrial Divide. New York: The Basic Books. Porter, Michael (1990) The Competitive Advantage of Nations. Basingstoke: Macmillan. —— (1998) ‘Clusters and the New Economics of Competition.’ Harvard Business Review, November/December, pp. 77–90. —— (2000) ‘Locations, Clusters and Company Strategies,’ in The Oxford Handbook of Economic Geography, G. L. Clark, M. Feldman, and M. Gertler Eds, Oxford: Oxford University Press, pp. 253–274. Pyke, Frank, Giacomo Becattini, and Werner Sengenberger (eds) (1990) Industrial Districts and Inter-firm Cooperation in Italy. Geneva: International Institute for Labor Studies/ILO. Rosenthal, Stuart S. and William C. Strange (2004) ‘Evidence on the Nature and Sources of Agglomeration Economies,’ in Handbook of Regional and Urban Economics, Volume 4, J. V. Henderson and J. F. Thisse Eds, Amsterdam: Elsevier B. V., pp. 2120–2171. Saxenian, AnnaLee (1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Cambridge, MA: Harvard University Press. Saxenian, AnnaLee (2001) ‘Bangalore: The Silicon Valley of Asia?’ Centre for Research on Economic Development and Policy Reform Working Paper No. 91, Stanford, CA: Stanford University. Saxenian, AnnaLee and Jinn-Yuh Hsu (2001) ‘The Silicon Valley-Hsinchu Connection: Technical Communities and Industrial Upgrading,’ Industrial and Corporate Change, Vol. 10, No.4, pp. 893–920. Storper, Michael and Anthony J. Venables (2005) ‘Buzz: Face-to-face Contact and the Urban Economy,’ in Clusters, Networks, and Innovation, Stefano Breschi and Franco Malerba Eds, Oxford: Oxford University Press, pp. 319–342. Sridharan, E. (2004) ‘Evolving Towards Innovation? The Recent Evolution and Future Trajectory of the Indian Software Industry,’ in India in the Global Software Industry: Innovation, Firm Strategies and Development, Authony P. D’Costa, and E. Sridharan Eds, Delhi: Macmillan India, pp. 27–50.

This page intentionally left blank

Part II Market Pooling and Organizational Change as Drivers for Creating Cluster and Innovation

This page intentionally left blank

7 A Comparative Analysis of Organizational Innovation in Japanese SMEs Generated by Information Communication Technology Masatsugu Tsuji and Shoichi Miyahara

7.1

Introduction

Small and medium-sized enterprises (SMEs) play an important role in the process of Japanese economic development in that they supply highquality parts to the manufacturing sector. In fact, the unsurpassed quality of Japanese products is largely due to SMEs. In this age of information, Japanese SMEs are required to face global challenges in order to survive. For this reason, they must utilize information and communications technology (ICT), which is key to the renovation of all business activities, especially organization. This chapter focuses on one category of innovation, namely the adoption of a new organizational structure originated by ICT use, and the analysis of factors that promote organizational innovation. Previous papers by Tsuji et al. (2005) and Bunno et al. (2006a, 2006b) attempted to determine what factors promote ICT use by SMEs. Field surveys, a mail survey and in-depth interviews were conducted in two of Japan’s most prominent SME clusters, which are located in Higashi-Osaka city in Osaka prefecture, and Ohta ward in the metropolitan area of Tokyo. In 2004, questionnaires were sent to more than 6,000 SMEs in the two clusters, which yielded nearly 1,200 replies. Although ICT use and innovative organizational renovation were not extensive in those SMEs, we also sent a mail survey to some of those chosen as the ‘top 100 SME business practices in the Kansai Area’ and ‘the 100 best SMEs, as selected by the Ministry of Economy, Trade and Industry (METI)’. Because these SMEs were thought to use ICT extensively, the same questionnaires were sent to them in December 2005. Of the 336 contacted, 137 replied. The results of this mail survey were summarized in Bunno et al. (2007). 231

232 Masatsugu Tsuji and Shoichi Miyahara

The questions sent to the SMEs dealt with (1) company characteristics (amount of capital, number of employees, etc.); (2) managerial orientation, which classifies SMEs as expanding, incentive-providing, adapting, or data-using; (3) business environment, such as the degree of competition; (4) purposes of ICT use, such as increasing profits and productivity; (5) expectations for ICT use; and (6) other factors such as ICT investment in the last fiscal year, and the company’s understanding of the importance of ICT in business management. The previous papers present common issues to be analysed, which are (1) identification of factors that promote ICT use by SMEs, with focus on management type and policies and (2) construction of an index to measure ICT use among SMEs. In these papers, the main issues are to construct a suitable index for ICT use by SMEs, and based on this, identify factors which promote ICT use. Tsuji et al. (2005) determined that the following items are good indicators of the degree of ICT use by SMEs: (1) the amount of software contributing to efficient utilization of managerial resources and (2) Internet use. Based on these data, they constructed an index in which the use of simple software or the Internet was worth 1 point, while more complicated and integrated utilization was worth 10. This scoring may seem somewhat arbitrary, but Bunno et al. (2006a) considered commonly used software or the Internet, which are used by many small SMEs, to be less important and thus worth fewer points. The points for each type of software were assigned according to the percentage of SMEs that use it, that is, the number of points is reciprocal to percent use. In other words, the more advanced and integrated the use, the more points were assigned to them. Bunno et al. (2006b, 2007) constructed an index to represent the degree of ICT use by SMEs by applying the Analytical Hierarchical Process (AHP).1 The index is mainly based on the utilization of (1) hardware and (2) information systems. The former consists of items such as (3) the number of PCs owned by an SME and (4) the number of PCs connected by networks such as LAN. The latter includes (5) software use, (6) Internet use and (7) security measures. In addition, the index takes into account (5) software use related to routine and non-routine work, (6) Internet use related to collecting and sending information as well as e-commerce and (7) security use related to technical and organizational measures for security. In order to calculate the AHP level, 11 ICT experts were asked to reply to questions about the importance of those indices and items. By utilizing these indices, Bunno et al. (2006a, 2006b, 2007) found that one of the most important factors was ‘ICT expectations’, such as ‘the restructuring of the whole business process’, which was identified as a significant factor in all of our estimations. SMEs that use ICT extensively believe in its effectiveness and invest a significant amount of money in it. It follows, then, that the most important way to promote ICT use among SMEs is to encourage them to be forward-thinking. Once they adopt such

Organizational Innovation in SMEs Generated by ICT

233

an outlook, SMEs can determine the exact ways in which they will introduce and use ICT to meet their specific goals. In addition, a probit analysis revealed that the behaviour of the CEO or top management was especially important. Since ICT use is a function of business management and strategy, the decisions made by senior managers are crucial. Even if SMEs operate under optimal conditions, they would not be able to use new technologies to their advantage unless their managers make correct decisions. We also identified policies that promote ICT investment by SMEs, such as tax and subsidy schemes as well as various deregulation measures. This chapter focuses on factors which promote organization innovation through ICT use by applying the index based on AHP developed by Bunno et al. (2006b). In addition, two groups, one comprising developed SMEs and the other underdeveloped, are compared in one model; the former is represented by SMEs in Higashi-Osaka/Ohta, while the latter by those selected by the IT Hyakusen Committee and METI (referred to as the IT Hyakusen group). Differences in the use of ICT for organizational innovation are analysed. In so doing, this chapter fully utilizes dummy variables to clarify the differences between two SME clusters; especially in addition to adding dummy variables to the constant term, we attach them to coefficients of independent variables. In the context of the Flowchart Approach to industrial clustering and innovation, this chapter rather relates to the latter, since two SME clusters we analyse here are already achieved the high level of industrial agglomerations and have been creating managerial as well as technological innovations based on agglomerations. Kuchiki and Tsuji (2004, 2008), and Tsuji et al. (2007) extensively analysed the agglomeration process currently occurring in East Asia and postulated a hypothesis which is referred to as the ‘Flowchart Approach’, which identifies factors which attract firms to the particular regions, which includes the following: (1) natural resources such as raw materials and human resources, skilled labour and professionals or unskilled labour; (2) physical infrastructures including highways, roads, airports, electricity, water supplies and other utilities; (3) social infrastructures such as legal, financial and intellectual property rights systems, deregulations, governmental institutional frameworks; and (4) incentive schemes provided by governments such as tax allowances and subsidies for investment and exports. The practical importance of the Flowchart Approach is to postulate the flowchart to achieve industrial clustering by checking whether the above conditions are satisfied or not. The Flowchart Approach to the endogenous innovation process, however, is not postulated yet, nor is the endogenous innovation process itself clearly defined yet.2 In order to construct this process, we have to grasp how industrial clustering develops into the innovation process, or more basically how innovations are created in the process of business activities. Fundamental data on how SMEs created organizational innovations by the appearance of ICT are presented

234 Masatsugu Tsuji and Shoichi Miyahara

in this chapter, and by making use of survey data this chapter also shows what are problems or obstacles for organizational innovation to SMEs, how SEMs struggle with ICT to promote efficiency of business and management, and how governmental policies support innovations. Based on these, the Flowchart Approach to the endogenous innovation process will be established. The chapter consists of six sections. In Section 7.2, we discuss the definition of organizational innovation generated by ICT use. The indices of ICT use by SMEs by focusing on AHP are presented in Section 7.3. The estimation variables employed to clarify the difference between the two SME groups are in Section 7.4. In Section 7.5, the estimation method and OLS, logit and probit estimations are presented, and factors to promote organizational innovation are identified based on survey responses. In Section 7.6, ICT use problems and policies encountered by SMEs are identified. Concluding remarks and suggestions for further analysis are given in the final section.

7.2 Organizational innovation and IT 7.2.1 Definition of innovation The Schumpeterian definition of innovation covers the following five changes: (1) new product; (2) new technology; (3) new materials; (4) new market; and (5) new form of management.3 In the context of innovation, focus has been on the first three factors; in other words, innovation has been discussed in physical technology terms. This is applicable to ICT (information and communication technology). ICT is an entirely new technology that has resulted in products such as PCs, servers and mobile communications tools. It has also played a part in the development of semiconductors and CPUs as new materials; namely, semiconductors are said to be the ‘source of industry’, just as steel used to be referred to as the ‘rice of industry’. ICT also created e-commerce, e-banking and online trading as a new form of market; now we can buy almost all commodities via the Internet. In some areas, the number of transactions taking place on the Internet far exceeds those through traditional trade. The development of business management has also increased greatly due to ICT. Business organizations have also been undergoing rapid transformation; ICT has created new forms of organization by destroying the traditional ones. The nature of ICT, (1) the ability to outsource work, (2) economy of speed and (3) economy of networking, has transformed the business model. Businesses can promote productivity by delegating their jobs to outside firms with higher productivity and professional skills, but they can obtain the same information as before through the ICT networks. ICT can also transmit large volumes of information within moments, which allows firms to make decisions quickly. Networks such as the Internet connect governments and research institutions such as universities, firms and

Organizational Innovation in SMEs Generated by ICT

235

individuals, which results in a dense flow of information. Under these circumstances, new ideas are more easily generated. 7.2.2 Case study: An SME supply chain that promotes exporting Although SMEs were once the primary force behind Japanese exporting, large firms eventually took over this role. However, with the spread of IT, the pendulum appears to be swinging back. This section examines an example of an export-oriented SME that has achieved organizational innovation by using ICT to help it construct its own international supply chain. 7.2.2.1 Profile of the firm Dan, a sock manufacturer, wholesaler and retailer, was established in 1968. It sells through its own shops in London as well as Japan. The company’s head office is located in Yao City, a suburb of Osaka. Its total capital is approximately ¥333 million, and it has 82 employees (since it has is fewer than 100 employees, it is classified as an SME). Sock manufacturers are divided into three types of firms: highly competitive national brands, specialized sock makers and SME sock makers. Dan’s socks go for ¥850 to ¥900 at its retail shops. Since most of its customers – mainly schoolgirls – usually make monthly visits to the store, Dan changes its stock every month. Customer information is collected through the firm’s POS system, which is directly connected to its distribution centre as well as its suppliers (sock knitters). This business model requires Dan to specialize in a wide range of designs and colours, although it produces relatively few of each item. Dan offers 500 items in 12 colours, for a total of 6,000 products. Management monitors sales at its shops and orders products on a weekly basis, so as to ensure that it can offer a full range of socks to appeal to its young customers. Unlike many Japanese SMEs that have outsourced production to countries such as China, Dan manufactures mainly in Japan. 7.2.2.2 Supply chain Dan’s president initially wanted to have factories located near the company’s outlets, but this proved unfeasible. However, the company achieved a similar result by using IT. It built its own supply chain system to transmit customer information through the POS system in real time, and to allow the factories, distribution centres and marketing departments to receive and utilize this information for decision-making. Dan has 40 knitters under contract, seven of which produce exclusively for the company. These knitters have from eight to 25 employees and are located close to the distribution centre – typically within a 10-minute drive. Sales information transmitted through the POS is received by the knitters, allowing them to update their own production plans. Dan has installed counters on its suppliers’ knitting machines, and production data are automatically transmitted to Dan’s managers, which allows them to monitor the production process. The total cost

236

Masatsugu Tsuji and Shoichi Miyahara

of constructing the distribution network and the supply chain network was ¥1.350 billion, most of which was provided through government subsidies. Dan’s supply chain is vertically structured, with Dan at the top, and the knitters below. Dan cannot organize upstream networks, such as those that distribute thread, since codes and purchase units vary from one company to another. This makes it impossible for Dan’s supply system to manage these transactions. Dan has unique purchasing and ordering schemes. Rather than ordering socks from its knitters, Dan requires the knitters to determine the amount of products they need to bring to the distribution centre using the information in the POS system. If products go unsold, the knitters must absorb the losses. This high degree of risk aversion regarding inventory precludes Dan from taking advantage of potential opportunities for large sales. After conducting a risk analysis, management chose to emphasize inventory management at the expense of potentially losing large orders. Although this marketing strategy could be criticized for being overly conservative, the company believes that it is a safe one for an SME. 7.2.2.3 Overseas shops and the international supply chain Dan is one of only three Japanese sock makers with overseas retail outlets. The company established Dan Socks, United Kingdom, in London in 2001, and opened its first shop in March 2002. It also sells socks through department stores such as Harrods. Dan’s overseas marketing strategy is different from those of other Japanese companies, which tend to rely on large trading firms for overseas sales it manages its overseas business directly. Prior to opening its London shop, Dan learned important skills from trading firms, including how to carry out tasks internally as much as possible and thereby reduce costs. The London shops are connected with the company’s home offices through the Internet-based POS system. The King Street shop has IBM computers, and the one on Neal Street has Dell computers. Both systems report data such as number of items sold, time of each sale, and customers’ gender and age, and can automatically calculate the value-added tax. All data are also transmitted to the knitters via Dan’s home offices. If additional socks are needed in London, the knitters can deliver them to the distribution centre at 24 hours’ notice. Once the customs declarations for export to the United Kingdom have been completed, the products are sent to Kansai International Airport. Although Dan tried to find suitable knitters in the United Kingdom, their quality did not meet the company’s standards. Due to differences between the two countries, such British regulations that prohibit the importation of assembled machinery from Japan, as well as voltage and safety standards, Dan gave up on its attempts to establish its own factory there. Thus, Dan ships all its products from Japan. The software for the POS system in the London shops was designed by six employees. Dan prefers to hire locally rather than outsource, despite the large cost differential. Although the ability

Organizational Innovation in SMEs Generated by ICT

237

to outsource to foreign manufacturers – one of the commonly cited advantages afforded by IT – purportedly allows firms to increase efficiency, such outsourcing also requires a substantial investment. Hence, Dan has found it more economical to subcontract to local companies in Japan.

7.3 Organization innovation in two groups of SMEs 7.3.1

Characteristics of the Higashi-Osaka/Ohta SMEs

Japanese manufacturing SMEs have supported the entire Japanese ‘Monostukuri (manufacturing)’ sector by supplying better parts, and the well-known superiority of Japanese products is based largely on the SMEs’ technological know-how and accumulated skills. In this chapter, two groups of SMEs, Higashi-Osaka/Ohta and IT Hyakusen, are compared in terms of organizational innovation initiated by IT. The former represents developed SMEs, and the latter developing ones. This chapter compares these two groups, and examines whether there are any differences in organizational innovation, and what kinds of factors affect the introduction of IT. Higashi-Osaka city and Tokyo’s Ohta ward were the objects of our field study, as these are the two largest SME clusters that have highly specialized technologies and networks of regional collaboration. The two regions, however, are not equivalent because the characteristics described below differ between them. SMEs in Higashi-Osaka manufacture completed products for the machinery and metalwork industries. More than 100 SMEs in HigashiOsaka manufacture unique products of their own, and maintain the largest shares of the markets for these products in Japan as well as abroad. Core sectors of SMEs located in Higashi-Osaka include metalware, plastics, electronics, general machinery and printing/publishing. Although they take contracts for some large ‘demand transporter’ companies such as Panasonic, Sanyo and Sharp, these SMEs tend to be more independent-minded and less focused on subcontracting than their counterparts in Ohta ward. In Higashi-Osaka, manufacturing SMEs have been constructing local networks through horizontal cooperation among SMEs who produce unique niche products and the accompanying peripheral products. In the Higashi-Osaka cluster, SMEs practice various cross-industrial exchanges so as to assimilate ideas for new technologies, product marketability, etc. These exchanges are strongly oriented towards creating novelty in the market. Most SMEs in Ohta Ward specialize in metalworking and processing, and are known to possess a high level of technical capability. Large as well as leading medium-sized companies in the electronic and automobile industries, such as Toshiba, Sony, NEC and Nissan have benefited by purchasing superior parts from them. Historically, large companies have chosen to locate in Tokyo’s metropolitan areas, which has allowed the SMEs in Ohta ward to develop strong ties and collaborations with them. This collaboration increases the SMEs’ effectiveness, but in turn, restricts their behaviour, that

238 Masatsugu Tsuji and Shoichi Miyahara

is, SMEs in Ohta ward tend to be passive and accept the role of subcontractor. Taking these differences between the two locations into consideration, the SMEs in Higashi-Osaka can be referred to as a ‘horizontal cluster’, while those in Ohta ward are a ‘vertical cluster’. We conducted field surveys, a mail survey and in-depth interviews in two of Japan’s most prominent SME clusters, located in Higashi-Osaka city, Osaka Prefecture, and Ohta ward, in the metropolitan area of Tokyo. In 2004, questionnaires were sent to more than 6,000 SMEs in the two clusters, which yielded nearly 1,200 replies. 7.3.2

Characteristics of IT Hyakusen SMEs

Higashi-Osaka and Ohta were both found to be less developed with regard to ICT use. To balance the pool, we included other SMEs that used ICT extensively, and the results for the two SME clusters were compared with those obtained by Bunno et al. (2006b). In so doing, we selected SMEs from among those chosen as the ‘top 100 SME business practices in the Kansai Area’ and ‘the 100 best SMEs, as selected by the Ministry of Economy, Trade and Industry (METI)’, referred to as the ‘IT Hyakusen Group’. The former SMEs were selected by the Kansai IT Strategic Committee according to their use of IT for management and business practice, and the latter SMEs were selected from all over Japan according to the same criteria. Some SMEs were selected for both groups. In December 2005, we sent a mail survey to 336 SMEs, including those mentioned above. Of the 336 contacted, 137 replied.

7.4 Index of ICT development 7.4.1 Index constructed by AHP Organizational innovation due to ICT use cannot be assessed with a single index because various factors are involved, including size, industry, business practices, etc. For the surveys, the following indicators of organizational innovation by SMEs were selected: (1) the number of PCs owned; (2) the number of PCs connected to networks such as LAN; (3) the extent to which software that contributes to the efficient utilization of managerial resources has been implemented; and (4) Internet use. No explanation is required for (1) and (2) since these indices are simple quantitative proxies for ICT use: the number of PCs is positively correlated with the degree of business activity transformation. Items (3) and (4) are more qualitative measures of ICT use, since having a large number of computers does not necessarily mean that they are being used efficiently. Initially, software packages – for example, those for accounting and marketing management – were introduced to promote efficiency for internal tasks. These applications were generally used on individual PCs, without any network connection. Businesses with more advanced ICT systems connect the users of these applications and make shared databases available to them. Item (3) sheds more light on this. Subsequently, the PCs in one or several locations may become connected to each other, generally through use of a groupware program.

Organizational Innovation in SMEs Generated by ICT

239

In previous studies by Tsuji et al. (2005) and Bunno et al. (2007), the index was constructed in such as way that questions 1 to 8 were worth 1 point, and questions 9 to 13 were worth 10 points. The description of ICT use in questions 1 to 8 was quite different from 9 to 13 since the latter dealt with more complicated and integrated utilization. This scoring may seem somewhat arbitrary. In this chapter, we utilize AHP, which is a more rigorous methodology for constructing an index. AHP was also adopted in the studies by Bunno et al. (2006a, 2006b, 2007), in which the steps of the decision-making process were assigned a numerical value. For example, when making a purchase, on what factors does a consumer base the decision? A consumer may consider the price, performance and design of various alternatives, before making a decision based on his/ her own criteria. AHP formulates the mechanism of such decision-making. It allows us to give a numerical value to the subjective parts of the decisionmaking process, which could be applied to a wide array of fields. Normally, each individual has not one but several evaluation criteria, and these often conflict with each other. In a consumer’s decision-making process, the ‘problem’ of what to choose comes first, followed by several ‘alternatives’. AHP attempts to objectify the decision-making process, assuming that there are some ‘criteria’ relating the specific ‘problem’ and the ‘alternatives’, using a hierarchical structure. In this chapter, the main factors that boosted ICT use among SMEs are broken down into the following two categories: establishment of hardware and utilization of the information system. The former has two sub-factors: (1) number of PCs owned and (2) number of PCs connected to networks such as LAN. The latter consists of following three: (3) the extent to which software that contributes to the efficient utilization of managerial resources has been implemented, (4) Internet use, and (5) security measures. Moreover, (3) software use includes that related to routine and non-routine work, and (4) Internet use includes that related to collecting and sending information as well as e-commerce. 7.4.2

Software and Internet use

The questions about software and Internet use that were included in the survey are explained in this section and listed in Tables 7.1 and 7.2, respectively. The number of questions in the survey made it impossible to ask pairwise questions for the determination of relative weights, which is basic to AHP. Therefore, we divided the questions into layers, which are indicated in Figure 7.1, in order to make the number of questions manageable. 7.4.3 Weight of items derived by AHP According to replies from 11 ICT experts, AHP divides the questions into three weighting layers (Figure 7.2). They rated ‘establishment and operation of an information system’ higher than ‘importance of hardware’; specifically, the former is 0.801, while the latter is 0.199. The former includes factors such as software use and Internet use (0.444 and 0.357, respectively).

240

Masatsugu Tsuji and Shoichi Miyahara Table 7.1

Question on software use Routine works

1. Sales management (including POS and bar code) 2. Accounting 3. Payroll management 4. Purchase management 5. Inventory management 6. Design management (include CAD/CAM) 7. Production management 8. Logistics Non-routine works 9. Enterprise resource planning (EPR) package 10. Customer relations management (CRM) 11. Group-ware (office information sharing system) 12. Sales force automation (SFA) 13. Supply chain management (SCM) Source: Authors

Table 7.2

Question on Internet use

Collection/exchange of information 1. Collection/exchange of information 2. PR of company and products 3. Efficient business management e-commerce 4. Net-banking 5. e-commerce with companies (BtoB) 6. e-commerce with consumers (BtoC) Source: Authors

Next, based on the AHP weighting, we calculated the degree of organizational innovation index for each SME and compared them with those of the two SME groups, Higashi-Osaka/Ohta and IT Hyakusen, which were selected by the IT Hyakusen Committee. The results are summarized in Table 7.3 and Figure 7.3. The average organizational innovation index value for IT Hyakusen and Higashi-Osaka/Ohta are 0.17 and 0.07, respectively, with the former using ICT at a more advanced level. Using this index, we can extract the essential factors promoting ICT use among SMEs. 7.4.4 Factors that affect organizational innovation Here, we explain variables that encourage organizational innovation caused by the development of ICT. The questionnaires asked SMEs about (1) company characteristics, (2) managerial orientation, (3) business environment,

Organizational Innovation in SMEs Generated by ICT Degree of organizational innovation

Improvement of hardware

1

Establishment and operation A of information system

Number of PCs per employee

1

Number of PCs connected to a LAN per employee

B

Software use

1

Routine works

1

Non-routine works

F

Internet use

Figure 7.1

241

C

Collections of information and PR

1

e-commerce

G

Layer of questions in AHP

Source: Authors

Degree of organizational innovation

Improvement of hardware

0.19898

Number of PCs per employee 0.04057 Number of PCs connected to 0.15841 a LAN per employee

Establishment and operation 0.80102 of information system

Software use

Routine works

0.44376

0.14088

Non-routine works 0.30288 Internet use

0.35726

Collections and PR 0.12524 e-commerce

0.23220

Figure 7.2 Weight obtained by AHP Source: Authors

(4) importance of the introduction of ICT, (5) expected results from ICT use, and (6) ICT investment in the last fiscal year. A list of these variables and related concrete questions is shown in Table 7.4.

Table 7.3

Index of organizational innovation of two groups Frequency

Degree of organizational innovation 0–0.05 0.05–0.1 0.1–0.15 0.15–0.2 0.2–0.25 0.25–0.3 0.3–0.35 Total Degree of organizational innovation IT Hyakusen Higashiosaka/Ohta Total Source: Authors

IT Hyakusen

Higashiosaka/Ohta

1 10 31 36 26 16 17

553 368 173 63 28 11 2

137

1,198

Average 0.39 0.13 0.16

Standard deviation 0.15 0.11 0.14

Ratio (%) Total 554 378 204 99 54 27 19 1,335

IT Hyakusen 0.73 7.3 22.63 26.28 18.98 11.68 12.24 100

Higashiosaka/Ohta

Total

46.16 30.72 14.44 5.26 2.37 0.92 0.17

41.5 28.31 15.28 7.42 4.27 2.02 0.22

100

100

Organizational Innovation in SMEs Generated by ICT

243

50 45 40

IT Hyakusen Higashiosaka /Ohta

35

Total 30 % 25 20 15 10 5 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Degree of organizational innovation Figure 7.3

Distributions of indices

Source: Authors

The first four of these six variables are explained as follows. First, (1) Company characteristics include variables such as amount of capital, number of regular employees, number of part-time employees, year of business establishment and the generation of the present owners. A detailed explanation is required to understand why (2) Management orientation is considered to be one variable. The questionnaires contain ten items on managers’ daily activities (Table 7.4). Since there is some overlap between the ten questions, an attempt was made to isolate the variables through component analysis. The respondents in the two groups, HigashiOsaka/Ohta and IT Hyakusen, created a pool of data from which the types of management were derived. Four variables, which account for 70.7 per cent of the total responses, were determined in this manner. The first of the questions determine the extent to which an SME is geared toward expansion as well as ICT training and education. This category is referred to as ‘orientation toward training/expansion’. The second category, ‘orientation toward information-sharing’ contains questions on the disclosure of business performance. The third variable includes questions on the extent to which firms learn from their mistakes, and on whether top management considers employees’ suggestions. Since such courses of action are indicative of management’s responsiveness, this factor is referred to as ‘orientation toward adapting’. The last variable, ‘orientation toward data use’, contains questions on how firms make use of data for decision-making. A summary of the statistics for all variables is shown in Table 7.5.4

Table 7.4

Result of component analysis Common factors

Management behaviour There is training and rotation to utilize each employee’s ability and knowledge The company offers ICT training to executives, managers and employees Employees are appraised of the company’s plan for next 2–3 years New lines of business are constantly being sought and products developed Company’s business performance is disclosed to employees Senior managers are provided broad responsibility and authorities Company studies competitors’ mistakes and learns from them Company listens to any employee’s opinion on how to improve management Past business data is extensively utilized in company’s management Monthly business data are utilized to improve management Eigenvalue Rotated factor pattern (%) Cumulative proportion (%) Source: Authors

Training/expansion type

Information-sharing type

Adaptive type

Data-using type

0.790

0.127

0.256

0.116

0.662

0.187

0.202

0.142

0.566

0.306

0.179

0.382

0.453

0.322

0.206

0.198

0.243

0.684

0.221

0.256

0.210

0.355

0.444

0.290

0.180

0.404

0.486

0.280

0.261

0.182

0.708

0.276

0.175

0.239

0.245

0.567

0.349

0.217

0.237

0.428

4.568 41.045 70.670

1.049 6.065

0.774 3.050

0.765 2.214

Table 7.5

Summary statistics Higashiosaka/Ohta

IT Hyakusen

Pooled data

Variables

Average

Standard deviation

Average

Standard deviation

Average

Standard deviation

Degree of organizational innovation * Characteristics Capital (10,000 yen) of firm The number of employees The number of part-time employees Operation years COE’s generation Managerial Training/expansion type behaviour ** Information-sharing type Apaptive type Data-using type Business We obtain new business environment partners every year The share of new products and services in our business is larger than before Many purchase orders are repeatedly from the same business partners We can price our own products In recent years, we have not been able to employ younger (30-year-old or younger) workers Importance of the introduction of ICT in business management

0.13 1,963.66 20.491 5.719

0.11 2,612.46 31.078 10.778

0.39 13,356.10 78.938 32.734

0.15 40,297.08 108.45 74.729

0.16 14,070 48.866 27.639

0.142 40,297 108.45 74.729

44.09 1.79 0 0 0 0 3.106

79.863 0.848 1 1 1 1 1.33

7.575 0.876 0 0 0 0 1.31

40.248 0.091 1 1 1 1 1.034

3.055

1.214

3.008

1.044

1.014

0.95

3.932

1.01

3.91

0.95

1.263

1.193

3.323 2.938

1.26 1.59

3.403 2.121

1.193 1.214

1.212 1.587

1.044 1.214

3.974

1.122

4.728

0.051

3.968

1.162

49.117647 40.248138 2.199 0.091 0 1 0 1 0 1 0 1 3.689 1.034

Continued

Table 7.5

Continued Higashiosaka/Ohta

Variables Expectation of ICT usage

Increased profit Higher productivity of routine works, such as administrative works Higher productivity of nonroutine business, such as project planning Higher speed of decision-making in management and business development Restructuring of the whole business process Active communication and accumulation sharing of information knowledge Precise understanding of customer needs Better customer satisfaction by improvement in services and products Company’s IT investment last fiscal year (10,000 yen) Amount of samples

IT Hyakusen

Standard deviation

Average

2.712 3.292

0.964 0.828

3.44 3.744

0.072 0.046

1.133 0.895

0.072 0.046

2.536

0.956

3.069

0.08

1.233

0.08

2.867

0.902

3.45

0.064

1.081

0.064

2.598

0.9

3.252

0.067

1.183

0.067

3.012

0.895

3.511

0.06

1.046

0.06

2.733

0.91

3.183

0.076

1.124

0.076

2.697

0.918

3.323

0.068

1.143

0.068

292.735 1,198

1,646.98

Average

2,781.20 137

Notes: * shows the result of Analytic Hierarchy Process. ** shows the four types of corporate management which was clarified by the method of the factor analysis. Source: Authors

Standard deviation

Pooled data

5,591.50

Average

1,574 1,335

Standard deviation

22,361

Organizational Innovation in SMEs Generated by ICT

7.5

247

Estimation

7.5.1 Estimation of the differences between two groups The factors that determine the particular scores obtained by each SME were examined (shown below) by using pooled data from the two groups. First, we examined whether there is a difference between the indices of the IT Hyakusen and Higashi-Osaka/Ohta groups by introducing a dummy variable that takes 1 for the former and 0 for the latter. The following regression model was constructed: Yi
b0  b1X1i  b2X2i  b3X3i  ... ...  bnXni  bdITdummy  ei,

(1)

where Yi is each SME’s index of organizational innovation; Xji denotes variables such as the characteristics of the SMEs, managerial behaviour, expectations for ICT use, etc.; bi indicates the coefficients to be estimated; IT dummy is a dummy variable attached to the IT Hyakusen group that uses bd as its coefficient; and ei is the residual. For this estimation, important variables are selected by checking a covariance matrix. The result of this estimation is summarized in Table 7.6. Table 7.6 indicates that that the IT Hyakusen dummy variable is significant at the 1 per cent level, and a significant difference exists between the average indices of organizational innovation of the two groups; namely, the average index of the IT Hyakusen Group is larger than that of the HigashiOsaka/Ohta Group by 0.28. Table 7.6

Result of OLS estimation

Variables Manufacturing Retail Capital We can determine prices Frequency of shipment of new products Recognition of ICT importance Improve profitability Efficiency of routine works Precise understanding of customer needs Amount of ICT investment Training/expansion type Adaptive type Data-using type Dummy variable attached to IT Hyakusen group Constants R2

Coefficient

t-value

0.009 0.019 0.016 0.006 0.005 0.02 −0.001 −0.005 −0.004 0.034 0.025 0.002 0.021 0.146 −0.022

1.338 1.401 4.382 *** 2.261 ** 1.756 * 6.352 *** −0.38 −1.291 −1.266 10.825 *** 5.975 *** 0.544 4.02 *** 12.389 *** −1.085

Note: ***, ** and * indicate the 1%, 5% and 10% significance levels. Source: Authors.

0.571

248

Masatsugu Tsuji and Shoichi Miyahara

7.5.2 OLS estimation of factors that affect indices of organizational innovation In this section, we identify the factors that affect the indices of organizational innovation in the two groups. In so doing, dummy variables are attached to each explanatory variable; again, 1 is taken for the IT Hyakusen group, and 0 for the Higashi-Osaka/Ohta group. As in the previous procedure, explanatory variables were selected by their degree of correlation with the indices. The variable coefficients estimated without the dummy variables indicated an effect on both groups, while those with dummy variables affected only IT Hyakusen group. The former is referred to as the ‘common effect’, and the latter as the ‘cross effect’. The following equation is used for estimation: Yi ⫽ b0 ⫹ ∑ bj X ji ⫹ ∑ b’(X j ji ITdummy i ) ⫹ b0 ’ ITdummy i ⫹ « i j =1

(2)

j =1

The results of the OLS estimation are given in Table 7.7. In this estimation, taking the effects on both groups into account, variables related to ‘Capital’, ‘Recognition of ICT importance’, ‘Amount of ICT investment’ and managerial behaviour-related variables such as ‘Training/expansion type’ and ‘Data-using type’ were found to be significant at the 1 per cent level, while ‘We can determine prices’ and ‘Precise understanding of customer needs’ were significant at the 10 per cent level. Regarding the cross effect-related to dummy variables, ‘Capita’ was significant at the 5 per cent level, while ‘We can determine the process’, ‘Frequency of shipment of new products’, ‘Efficiency of routine work’, and ‘Training/expansion type’ were significant at the 10 per cent level. In order to identify the factors in more detail, we classified the variables used for estimation into four categories: Groups I, II, III and IV. Group I variables have a significant cross effect (affect only IT Hyakusen) as well as a common effect (affect both groups); Group II variables have only a cross effect, in other words, they affect only the IT Hyakusen group; Group III variables affect both groups; and Group IV variables are not significant and can be ignored. This classification scheme is shown in Table 7.8. The variables in Group I, such as ‘Capital’, ‘We can determine prices’ and ‘Training/expansion type’ were significant to both groups, but affected the IT Hyakusen group more (except for ‘Capita’, which had a negative coefficient). This implies that larger SMEs in general tend to make more advanced organizational innovations, but for SMEs with developed ICT, such as those in the IT Hyakusen group, the amount of capital is less relevant to innovation. For Group II, ‘Efficiency of routine work’ and ‘Frequency of shipment of new products’ were associated only with the IT Hyakusen group. Well-developed SMEs ship new products to the market more frequently and are eager to seek efficiency through organizational innovation. The Group III variables ‘Recognition of ICT importance’, ‘Data-using type’, ‘Precise understanding of customer need’ and ‘Amount of ICT investment’ were common characteristics of both IT Hyakusen and Higashi-Osaka/Ohta groups.

Organizational Innovation in SMEs Generated by ICT Table 7.7

249

Factors affecting organizational innovation (1)

Variables

Coefficient

t-value

Manufacturing Retail Capital We can determine prices Frequency of shipment of new products Recognition of ICT importance Improve profitability Efficiency of routine works Precise understanding of customer needs Amount of ICT investment Training/expansion type Adaptive type Data-using type Dummy variable attached to IT Hyakusen group

0.012 0.002 0.023 0.004 0.005 0.02 −0.001 −0.006 −0.006 0.033 0.021 0.002 0.021 −0.196

1.641 0.138 5.52 *** 1.703 * 1.544 6.243 *** −0.394 −1.512 −1.784 * 9.814 *** 4.987 *** 0.547 4.038 *** −1.636

Cross-effect (dummy X variable) Manufacturing Retail Capital We can determine prices. Frequency of shipment of new products Recognition of ICT importance Improve profitability Efficiency of routine works Precise understanding of customer needs Amount of ICT investment Training/expansion type Adaptive type Data-using type Constant R2

−0.026 0.025 −0.022 0.017 0.018 0.022 −0.006 0.035 0.013 0.005 0.026 0.028 −0.01 −0.019

−1.232 0.801 −2.374 ** 1.877 * 1.768 * 1.224 −0.444 1.851 * 1.026 0.564 1.653 * 1.407 −0.443 −0.936 0.591

Note: ***, ** and * indicate the 1%, 5% and 10% significance levels. Source: Authors.

7.5.3 Probit/logit estimation of factors that affect the innovation index The variables of logit and probit estimations were constructed using data obtained through the mail survey, and usually take discrete value, which make them better than OLS. In these estimations, the SMEs were divided into two categories, ones with a larger index than the average, and ones with smaller. The equations to be estimated are as follows: Logit model: F( xi⬘␤ )⫽

exp( xi⬘␤ ) 1⫹exp( xi⬘␤ )

(3)

250 Masatsugu Tsuji and Shoichi Miyahara Table 7.8

Factors affecting organizational innovation (2)

Variables

Common effect

Cross effect

Coefficient

Coefficient

0.023 Capital We can determine prices I Training/expansion type

−0.022 0.004 0.017 0.021 0.026

Efficiency of routine works II

0.018 Frequency of shipment of new products 0.035 Recognition of ICT importance 0.02 Data-using type

III Precise understanding of customer needs Amount of ICT investment

0.021 −0.006 0.033

Manufacturing Retail IV Improve profitability Adaptive type Notes: I: Variables which cross and own effects are significant. II: Variables which only cross effects are significant. III: Variables which only own effects are significant. IV: Not significant at all. Source: Authors.

Probit model: F( xi⬘␤ )⫽⌽( xi⬘␤ )

(4)

where F denotes the standard normal distribution function, and the xi variables are similar to those in the OLS estimation. The results of the Logit and Probit estimations are shown in Table 7.9. They are similar to the OLS estimation in the previous section. ‘Amount of capital’ (marginal effect: 0.0572 and 0.0656), ‘ICT investment of previous year’ (marginal effect: 0.6417 and 0.3728), ‘Data-using type’ (marginal effect: 0.0760 and 0.0789), and IT Hyakusen dummy (marginal effect: 0.06821 and 0.2984) were all significant at the 1 per cent level. ‘Recognition of ICT importance’ (marginal effect: 0.0399 and 0.0418) was significant at the 5 per cent level. In addition, ‘We can determine prices’ (marginal effect: 0.0247 and 0.0258) and ‘Training/expansion type’ (marginal effect: 0.0431

Table 7.9

Probit/logit estimation Logit-model

Variables

Coefficient

z-value

Manufacturing Retail Capital We can determine prices Frequent shipment of new products Recognition of ICT importance Improve profitability Efficiency of routine works Precise understanding of customer needs Amount of ICT investment Training/expansion type Adaptive type Data-using type Dummy variable attached to IT Hyakusen group Constants Log livelihood

0.2712557 0.7308473 0.3140201 0.1197586 0.0974063 0.1934601 0.1310588 −0.0212161 −0.0685335

1.36 1.56 2.9 *** 1.71 * 1.26 2.04 ** 1.29 −0.19 −0.71

0.6416951 0.2087655 0.0059616 0.3684085 2.177614 −2.937576

6.3 *** 1.87 * 0.05 2.63 *** 3.55 *** −4.99 *** −412.66702

Note: ***, ** and * indicate the 1%, 5% and 10% significant levels. Source: Authors

Probit-model

Marginal effect

Coefficient

z-value

Marginal effect

0.0572834 0.1289557 0.0648077 0.0247159 0.0201028 0.0399265 0.027048 −0.0043786 −0.014144

0.1532331 0.4303353 0.1873305 0.0738191 0.060102 0.1193981 0.079551 −0.0160118 −0.0316916

1.28 1.59 3 *** 1.76 * 1.3 2.07 ** 1.31 −0.23 −0.55

0.054518 0.1338354 0.065617 0.0258569 0.0210522 0.041822 0.0278646 −0.0056085 −0.0111007

0.1324336 0.0430852 0.0012304 0.0760324 0.2934461

0.3727549 0.1289496 −0.0001967 0.2278896 1.185455

−1.786065

−5.03 *** −412.13722

6.64 *** 1.93 * 0 2.72 *** 3.93 ***

0.1305662 0.0451677 −0.0000689 0.0798237 0.2983793

252

Masatsugu Tsuji and Shoichi Miyahara

and 0.0452) was significant at the 10 per cent level. These results are consistent with those of the OLS analysis.

7.6 Problems and policies for organizational innovation 7.6.1 Problems with organizational innovation using ICT Thus far, analysis has focused on factors that encourage organizational innovation through ICT use. This section examines innovations in business organization and ICT-related issues that SMEs are facing, so as to identify problems and recommend policies that could be implemented to solve them. The kinds of obstacles faced by the SMEs in each category of the developed (IT Hyakusen) and the developing (Higashi-Osaka/Ohta) groups were also examined. The obstacles are summarized in Q9 of the questionnaire. We utilized the same analysis as in the previous section, with the addition of two kinds of dummy variables: the IT Hyakusen dummy was used as a constant, and each variable; the latter is also referred to as the cross effect. The results of the OLS estimation using equation (2) are shown in Table 7.10. A list of the independent variables can be found in Table 7.11. The information in Table 7.10 was used to identify the issues for organizational innovation. Since the dependent variable is the degree of organizational innovation and the dependent variables the kinds of obstacles, coefficients are expected to be negative, that is, because of these serious obstacles, SMEs have been reluctant to introduce ICT. The positivity of coefficients, on the other hand, can be interpreted in such way that even if SMEs have these problems, they make effort to introduce ICT in order to solve them. In this sense, the negativity (positivity) of coefficients implies negative (positive) reasons of decision-making towards ICT investment. As in the previous section, related variables are classified into four categories by significance. The Group I variables, ‘Unclear objectives of management’ and ‘ICT security is a major concern’ are significant to both the common and cross effects. Care should be taken to the following: coefficients of the cross effect are negative, while those of the common effect positive. Since the latter shows those of Hogashi-Osaka/Ohta SMEs, while those of IT Hyakusen are expressed by totals of two effects, that is, IT Hyakusen SMEs have negative coefficients with respect to these two variables.5 The opposite signs of these variables can be interpreted as follows: IT Hyakusen SMEs are already achieved the higher level of ICT use, and these two might not be major reasons to introduce ICT, while Higashi-Osaka/Ohta SMEs are relatively at the low level of ICT use and they strongly concern these problems and thus have been positively introducing ICT.6 The second category was more significant for the IT Hyakusen group: ‘Employees’ lack of ICT knowledge’, ‘Introduction to ICT is left up to the hardware/software makers’, and ‘Lack of workers’ cooperation with ICT usage at the office’. The first two have negative signs, which implies that they are less serious issues for the IT Hyakusen group, but the last one is positive; this problem is more important

Organizational Innovation in SMEs Generated by ICT Table 7.10

253

Problems of organizational innovation by SMEs (1)

Variables

Coefficient

Common effect Lack of leadership regarding ICT use Unclear objectives of management ICT has been introduced without any restructuring of works Lack of employees who can use ICT Lack of workers’ ICT knowledge Lack of workers’ cooperation with ICT usage at the office Lack of ICT advisers We leave everything of ICT introduction to ICT adviser(s) We leave everything of ICT introduction to ICT makers Lack of software that we need We can’t keep up with technological innovation Each business partner wants to adopt its own ICT systems ICT investment does not yield explicit profit ICT investment is very costly We have deep concern for information security, if ICT is introduced It takes time to introduce ICT Others IT Hyakusen dummy variable Cross-effect (dummy X variable) Lack of leadership regarding ICT use Unclear objectives of management ICT has been introduced without any restructuring of works Lack of employees who can use ICT Lack of employees’ ICT Knowledge Lack of workers’ cooperation with ICT usage at the office Lack of ICT advisers We leave everything of ICT introduction to ICT adviser(s) We leave everything of ICT introduction to ICT makers Lack of software that we need We can’t keep up with technological innovation Each business partner wants to adopt its own ICT systems ICT investment does not yield explicit profit ICT investment is very costly We have deep concern for information security, if ICT is introduced It takes time to introduce ICT Others Constants R2 Note: ***, ** and * indicate the 1%, 5% and 10% significance levels. Source: Authors.

t-value

−0.025 0.014 0.015

−3.139 *** 1.696 * 1.694 *

0.007 0.01 0.007 −0.008 0.031

0.887 1.302 0.551 −0.998 2.177 **

0.009

0.834

0.017 −0.026 0.046 −0.013 0.022 0.051

2.025 ** −2.917 *** 4.895 *** −1.626 2.872 *** 6.831 ***

−0.024 −0.032 0.302

−2.29 ** −1.655 * 15.589 ***

−0.013 −0.043 −0.029

−0.391 −1.676 * −1.028

−0.013 −0.06 0.095

−0.527 −2.315 ** 1.763 *

0.033 −0.085

0.928 −1.45

−0.1

−2.101 **

0.026 0.042 −0.043

0.914 1.303 −1.462

−0.001 −0.01 −0.06

−0.019 −0.424 −2.764 ***

0.015 0.11 0.107

0.313 2.488 ** 18.94 *** 0.39

254 Masatsugu Tsuji and Shoichi Miyahara

for them. The eight variables in Group III are issues common to both groups. A positive (negative) sign implies that this variable is more related to SMEs with larger (smaller) innovation indices. ‘ICT has been introduced without any restructuring of work’, ‘Introduction of ICT is left to the ICT adviser(s)’, ‘Lack of necessary software’, ‘Each business partner wants to adopt its own ICT systems’ and ‘ICT investment is very costly’ are more serious to SMEs with larger indices. On the other hand, ‘Lack of leadership regarding ICT use’, ‘We can’t keep up with technological innovation’ and ‘It takes time to introduce ICT’ are more important problems for SMEs with smaller indices. The last three variables in particular seem to be common hurdles for the introduction of ICT into small SMEs. Table 7.11

Problems of organizational innovation by SMEs (2) Common effect Variables

I

II

Coefficient

Cross effect Coefficient

Unclear objectives of management

0.014

−0.043

We have deep concern for information security, if ICT is introduced. Others Lack of employees’ ICT Knowledge

0.051

−0.085

−0.032

0.11 −0.06

Lack of workers’ cooperation with ICT usage at the office We leave everything concerning ICT introduction to ICT makers Lack of leadership regarding ICT use ICT has been introduced without any restructuring of works We leave everything concerning ICT introduction to ICT adviser(s)

III Lack of software that we need We can’t keep up with technological innovation. Each business partner wants to adopt its own ICT systems. ICT investment is very costly It takes time to introduce ICT Lack of employees’ who can use ICT IV Lack of ICT advisers ICT investment does not yield explicit profit Notes: I: Variables which cross and own effects are significant II: Variables which only cross effects are significant III: Variables which only own effects are significant IV: Not significant at all. Source: Authors.

0.095 −0.1 −0.025 0.015 0.031 0.017 −0.026 0.046 0.022 −0.024

Organizational Innovation in SMEs Generated by ICT

255

The problem common to both groups is ‘Lack of leadership regarding ICT use’, which indicates that Japanese SMEs still need top management with strong ICT leadership. The surveys revealed that both ICT knowledge for employees and ICT leadership for top management need to be improved, which implies that the problem is related to human resources. This is key to policies geared toward promoting ICT in SMEs. 7.6.2 Policies advantageous to innovation as suggested by empirical research This section analyses the kinds of policies that are required to encourage adoption of organizational innovations. In order to examine this problem, the OLS model (equation (2)) was used. The results of our estimations are shown in Tables 7.12 and 7.13. Table 7.12

Policy desired for organizational innovation (1)

Variables Common effect Opening of ICT seminars Implementation of education for PC operation Adviser system Low-interest loans for ICT Low-interest lease for ICT Tax exemptions on ICT investment Support for opening new portals Deregulation Commendation of small company business models that make use of ICT Introduction of e-bidding system Others Dummy variable attacked to IT Hyakusen Cross-effect (dummy X variable) Opening of ICT seminars Implementation of education for PC operation Adviser system Low-interest loans for ICT Low-interest lease for ICT Tax exemptions on ICT investment Support for opening new portals Deregulation Commendation of small company business models that make use of ICT Introduction of e-bidding system Others Constants R2

Coefficient

t-value

0.007 −0.002 0.005 0.016 0.016 0.064 −0.002 0.039 0.059

0.969 −0.205 0.656 2.09 ** 2.199 ** 8.887 *** −0.132 4.469 *** 3.546 ***

0.031 0.012 0.277

2.482 ** 0.834 15.199 ***

−0.013 −0.072 −0.046 −0.019 −0.004 −0.042 0.061 0.009 −0.073

−0.476 −2.082 ** −1.764 * −0.736 −0.152 −1.943 * 1.405 0.378 −2.453 **

0.119 0.077 0.088

Notes: ***, ** and * indicate the 1%, 5% and 10% significance levels. Source: Authors.

2.74 *** 2.053 ** 15.795 *** 0.404

256 Masatsugu Tsuji and Shoichi Miyahara Table 7.13

Policy desired for organizational innovation (2)

Variables I

II

III IV

Common effect

Cross effect

Coefficient

Coefficient

0.064 0.059

−0.042 −0.073

0.031

0.119 −0.072

Tax exemptions on ICT investment Commendation of small company business models that make use of ICT Introduction of e-bidding system Implementation of education for PC operation Adviser system Others Low-interest loans for ICT Low-interest lease for ICT Deregulation Support for opening new portals Opening of ICT seminars

−0.046 0.012 0.016 0.016 0.039

Notes: I: Variables which cross and common effects are significant. II: Variables which only cross effects are significant. III: Variables which only common effects are significant. IV: Not significant at all. Source: Authors.

‘Tax exemptions for ICT investments’, ‘Grants and other financial support for ICT investments’, ‘Commendation of small company business models that make use of ICT’, ‘Introduction of an e-bidding system’, ‘Lowinterest loans for ICT’ and ‘Low-interest leases for ICT’ and ‘Deregulation’ were policies desired by both SME groups, and were positively related to the innovation index. The IT Hyakusen group was particularly concerned about the e-bidding system because it is related to a higher rate of ICT use. In contrast, they are less interested in subsidies, such as tax exemption, and training and education. These are consistent within the IT Hyakusen group, since they have already achieved a certain level of organizational innovation.

7.7

Conclusions

Based on the intensive mail surveys conducted in two Japanese major SME groups, Higashi-Osaka/Ohta ward and IT Hyakusen SMEs, the latter of which was recognized for their ICT use. The data collected by the survey were used to create an index of ICT use by these SMEs, thereby clarifying the factors that promote SME organizational innovation via ICT use. SMEs that use ICT intensively were found to believe in its effect and invest a lot of money

Organizational Innovation in SMEs Generated by ICT

257

in it so as to improve their businesses. It follows from this that the most important way to promote ICT use among SMEs is to encourage them to be forward-thinking. Once they adopt such an outlook, they can determine the exact ways in which they will introduce and use ICT to meet their specific goals. The estimation of the problems related to ICT introduction, which was particularly relevant to the Higashi-Osaka/Ohta SMEs, showed a positive relationship between ICT index and ‘Each business partner wants to adopt its own ICT systems’ or ‘Information security is a major concern’. This indicates that the issues are mainly location- and human resource-related due to the following reasons: ● ●

● ●

Large firms want subcontractors to use the firms’ ICT systems. A large amount of money is required to comply with the ICT demands of these large firms. There is a lack of human resources to handle ICT. There are security concerns for data related to customers, transactions and privacy.

SMEs with advanced levels of ICT use in these regions tend to shift all of their business activities or solve managerial problems by restructuring their businesses. In so doing, problems involving customer relationships and ICT utilization by employees become important. IT Hyakusen SMEs, on the other hand, introduce and operate ICT without help from outside experts, but through their own employees. In addition, they do not introduce it in such a way so as to restructure all business activities, but rather to improve their businesses gradually. In this way, they improve employee ICT capability and renovate ICT systems. This chapter focuses on extracting factors that promote ICT use by SMEs. Once identified, they can be used to establish suitable policy measures. Our investigations revealed that Higashi-Osaka/Ohta requested tax exemptions and subsidies for ICT investment, which indicates that shortage of funds is the most serious obstacle for ICT development. This problem is doubled because of the need for ICT updates as technology improves. IT Hyakusen SMEs, on the other hand, are interested in measures that expand their business chances, such as the introduction of an e-bidding system. The findings in this study are needed as a basis for further policy measures. Many policies have been implemented by various ministries of the government so far (Tsuji et al. 2005; Small and Medium Enterprise Agency 2004; 2001, 2002, 2003, 2004), but they can hardly be considered successful. Proper policy measures based on rigorous research are needed, but have yet to be established.

258

Masatsugu Tsuji and Shoichi Miyahara

Appendix Questionnaires Please write your name and contact information Company name Address

 ⳕ

Telephone number

FAX number

Name of replier

Title/position

E-mail

Section A: IT in your company IT (Information Technology) refers to information equipment or information technology, such as personal computers and the Internet. In this section, we would like to ask about your company’s initiatives for the renovation of business management by the utilization of IT. Q1) How many PCs (including those on lease or rental) does your company have?

Q1–1) How many PCs among them are connected to a LAN (local area network)?

Q2) What kind of software does your company use? And, what kind of software not currently in use would you like to use in the future? Using

Would like to use

1. Sales management (including POS and bar code)

1

1

2. Accounting

2

2

3. Payroll management

3

3

4. Purchase management

4

4

5. Inventory management

5

5

6. Design management (include CAD/CAM)

6

6

7. Production management

7

7

8. Logistics

8

8

9

9

10. Customer relations management (CRM)

9. Enterprise resource planning (EPR) package

10

10

11. Group-ware (office information sharing system)

11

11

12. Sales force automation (SFA)

12

12

13. Supply chain management (SCM)

13

13

14. Others

14

14

Organizational Innovation in SMEs Generated by ICT

259

Q3) Does your company use the Internet? (Choose one) 1 Yes

2 No

↓ ↓ If ‘Yes’, please go to Q3–1 through Q3–4. If ‘No’, please skip to Q4. Q3–1) When did your company start to use the Internet? Year

Q3–2) What kind of the Internet access are you using? (Check one or more) 1. Leased circuit 2. Fibre optical 3. Cable modem

4. Asymmetric Digital Subscriber Line (ADSL) 5. Integrated Services Digital Network (ISDN) 6. Dial-up

Q3–3) What purposes describe the current Internet use of your company? (Check one or more) 1 Development of the company’s home page or other related web pages. 2 Employees’ personal e-mail addresses (number of such addresses). 3 Electronic boards, and/or electronic meeting boards. 4 Utilizing mailing lists of customers and business partners. 5 Own domain name. 6 Others.

Q3–4) What is the purpose of your Internet use? 1. Collection /exchange of information 4. Efficient business management 7. Others 2. PR of company and products 5. e-commerce with companies (BtoB) 3. Net-banking 6. e-commerce with consumers (BtoC)

If you checked ‘e-commerce’ (5 or 6) at Q3–4, please go to Q3–5 and Q3–6. If not, please skip to Q4. Q3–5) How was your company’s e-commerce performance in the last fiscal year? Ratio of sales by the internet (total sales is 100%) % About Compared with that of three years ago, this ratio is 1. Increased 2. Almost same 3. Decreased

Ratio of purchase through the Internet (All purchase = 100%) About %

260 Masatsugu Tsuji and Shoichi Miyahara

Compared with that of three years ago, the ratio is 1. Increased 2. Almost same 3. Decreased

Q3–6) The proportion of business with the companies in your region conducted by e-commerce, compared with that of three years ago, is: 1. Increasing 2. Almost same 3. Decreasing

Q4) What kind of IT training does your company provide to employees? (Check one or more) 1. Support participation in outside IT training and seminars 2. In-house IT training and seminars

3. Support individual 5. Others learning 6. We do not have any 4. Employ persons with high IT ability

Q5) What kind of security measures does your company have? And, what kind of security does your company plan to introduce? (Check one or more) We have We’d like to have 1. 2. 3. 4. 5. 6. 7. 8. 9.

Principle of security Risk analysis Ranking of classified information Prohibition of leaking classified information and customer information Control of passwords Introduction of firewalls Anti-virus measures System audition and information security audition Others

1 2 3 4 5 6 7 8 9

1 2 3 4 5 6 7 8 9

Section B: Your corporate opinion about IT use

We expect not so much

We expect not at all

We expect to some extent

We expect very much

Q6) What benefits of IT use do you expect to enjoy in the following areas?

1. Increased profit

1

2

3

4

2. Higher productivity of routine works, such as administrative works

1

2

3

4

3. Higher productivity of non-routine business, such as project planning

1

2

3

4

Organizational Innovation in SMEs Generated by ICT

261

4. Higher speed of decision-making in management and business development

1

2

3

4

5. Restructuring of the whole business process

1

2

3

4

6. Active communication and accumulation sharing of information knowledge

1

2

3

4

7. Close cooperation with customers and business partners

1

2

3

4

8. Precise understanding of customer needs

1

2

3

4

9. Better customer satisfaction by improvement in services and products

1

2

3

4

We expect very much

We expect to some extent

We expect not so much

We expect not at all

Q7) How satisfied are you with the current benefits of your company’s IT use in the following areas? If your company has not used IT so far, please skip to Q8.

1. Increased profit

1

2

3

4

2. Higher productivity of routine works, such as administrative works

1

2

3

4

3. Higher productivity of non-routine business, such as project planning

1

2

3

4

4. Higher speed of decision-making in management and business development

1

2

3

4

5. Restructuring of the whole business process

1

2

3

4

6. Active communication and accumulation sharing of information knowledge

1

2

3

4

7. Close cooperation with customers and business partners

1

2

3

4

8. Precise understanding of customer needs

1

2

3

4

9. Better customer satisfaction by improvement in services and products

1

2

3

4

Q8) Do you think there are any problems with your company’s current IT use? (Check one) 1. We have serious problem(s). 2. We have some problem(s).

3. Uncertain

4. We have few problems. 5. We have no problems at all.

262

Masatsugu Tsuji and Shoichi Miyahara

Q9) What is the problem(s) in your company’s current IT use? (Check one or more) 1. 2. 3. 4. 5. 6. 7. 8. 9.

Nobody takes leadership for IT use. No corporate target for IT use has been clarified. IT has been introduced without any restructuring of works. Shortage of staff with strong IT skills. Employees’ poor IT knowledge. Lack of workers’ cooperation with IT use at the office. Lack of appropriate IT advisers. The company leaves adviser(s) to introduce IT as they like. The company leaves vendors (manufacturers) to plan and introduce IT as they like. 10. There is no software applicable to our business and works. 11. We cannot catch up with the rapid development of IT. 12. Each business partner wants to adopt their own IT systems. 13. IT investment does not yield explicit profit. 14. IT investment is very costly. 15. We have deep concern for information security, if IT is introduced. 16. We have deep concern for leakage of personal data, if IT is introduced. 17. IT introduction takes too much time. 18. Others (please specify: )

Q9–1) What are the three most serious among the above 18 problems. #1

#2

#3

Q10) How important is the introduction of IT in business management? (Check one) 1. Very important 2. Somewhat important

3. Uncertain

4. Not very important 5. Not important at all

Q11) How much does your company plan to use IT more? (Check one) 1. Very much 2. To some extent

3. Uncertain

4. Not so much 5. Not at all

Q12) How much does your company plan to increase IT investment? (Check one) 1. Very much 2. To some extent

3. Uncertain

4. Not so much 5. Not at all

Q13) How much was your company’s IT investment last fiscal year? About

yen

Organizational Innovation in SMEs Generated by ICT

263

Section C: Your company’s corporate policy Q14) What are the current important challenges for your company? (Check one) 1. To secure orders and sales 2. To secure positive profit and fund management 3. To secure good employees and human resource management 4. Development of ability for planning, R&D and technology

5. IT utilization 6. Company succession 7. Other (please specify: )

Very much

To some extent

Uncertain

Not so much

Not at all

Q15) How applicable are the following statements to your company?

1. Company’s business performance is disclosed to the employees.

1

2

3

4

5

2. Past business data is utilized in business management.

1

2

3

4

5

3. The executives are provided with broad responsibility and jurisdiction.

1

2

3

4

5

4. We study other companies’ failures and learn from them.

1

2

3

4

5

5. We hear any employee’s opinion concerning better management.

1

2

3

4

5

6. We are constantly developing new business and products.

1

2

3

4

5

7. Monthly business statistics are utilized for management.

1

2

3

4

5

8. We offer IT training to corporate executives, managers and workers.

1

2

3

4

5

9. Training and personnel job rotation are conducted so as to mobilize each employee’s ability and knowledge.

1

2

3

4

5

10. Employees are well informed of the 2- or 3-year future direction of the company.

1

2

3

4

5

11. We obtain new business partners every year.

1

2

3

4

5

12. Many purchase orders are repeatedly from the same business partners.

1

2

3

4

5

264 Masatsugu Tsuji and Shoichi Miyahara

13. We can price our own products.

1

2

3

4

5

14. Competition with rival companies has recently become more severe.

1

2

3

4

5

15. New entrants from other business fields have recently increased in our market.

1

2

3

4

5

16. The share of new products and services in our business is larger than before.

1

2

3

4

5

17. In recent years, we have not been able to employ younger (30-year-old or younger) workers.

1

2

3

4

5

Section D: Future IT and policy Q16) What future government policies regarding IT use would you like to see implemented? (Check one) 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

IT seminars Training for PC use Training for website development Advice for IT promotion Low-interest loans for IT Lease of IT with low fees Tax exemption incentives for IT investment Grants and financial supports for IT investment related projects Support for opening new portal sites Deregulation Commendation of small company business models that make use of IT e-procurement, e-purchase Others (please specify: )

Q16–1) What are the three most expected among the above 13 initiatives? #1

#2

#3

Q17) The Japanese government is promoting deregulations by establishing a ‘Special Zone for the Structural Reform’. The ‘Special Zone for IT and new business promotion’ is described as follows. Please give us your opinion: ‘The Special District for IT and New Business Promotion’ is designed for the creation of new industries, the invitation and incubation of new businesses, and better services for the public through the development of IT and telecommunications infrastructure based on fibre optical networks and deployed by local governments. This deregulation makes it much easier to conduct businesses with IT use in various fields.

To some extent

Uncertain

Not so agreeable

Not agreeable at all

265

Quite agreeable

Organizational Innovation in SMEs Generated by ICT

1. The Special Zone is a large advantage for Japanese manufacturers to achieve stronger international competitiveness.

1

2

3

4

5

2. We expect the Special Zone to provide us with financial support.

1

2

3

4

5

3. We expect the Special Zone to support our human resources development.

1

2

3

4

5

4. Only large companies or IT industries will be able to take advantage of the Special Zone, and other SMEs may take nothing from it.

1

2

3

4

5

5. If the Special Zone is established in Higashiosaka City or Ohta-ku (Tokyo), we’d like to make practical use of it.

1

2

3

4

5

Q17–1) If the above described ‘Special District for IT and New Business Promotion’ is established in Higashiosaka City or Ohta-ku (Tokyo), what do you expect of it?

Q18) Would your company like to promote collaborative projects with research institutions of universities and other organizations? (Check one) 1. Yes, very much 2. Yes, to some extent

3. Uncertain

4. Not really 5. Not at all

If you check Yes (1 or 2), please go to Q18–1. If you check 3 to 5, please skip to Q19). Q18–1) In what specific field(s) would you like to collaborate?

Q19) Do you consider IT communication spaces (ex. Internet) to fundamentally be public spaces or private spaces? (Check one) 1. Public spaces, absolutely 2. Primarily public spaces

3. Uncertain

4. Primarily private spaces 5. Private space absolutely

266

Masatsugu Tsuji and Shoichi Miyahara

Don’t agree at all

1

2

3

4

5

2. Data obtained by IT is fully open to the public and businesses.

1

2

3

4

5

3. IT achieves higher autonomy of people and business.

1

2

3

4

5

4. IT will widen the gap between large companies and SMEs.

1

2

3

4

5

5. IT will widen the economic gap between the developed and developing countries.

1

2

3

4

5

6. IT promotes the globalization of the world.

1

2

3

4

5

7. IT creates new markets and replaces old ones.

1

2

3

4

5

8. IT makes it possible to have responses from the world to what we disseminate, so that we can adjust direction anytime.

1

2

3

4

5

Uncertain

1. IT provides all people with equal opportunity for information exchange.

Quite agree

Not so agreeable

Agree to some extent

Q20) How much do you agree with the following views? (Check one for each view)

Q21) Any further opinion and request regarding IT and new business will be appreciated.

Section E: Corporate profile Q22) Established Year: Month:

Present CEO is: 1. Founder of the company 2. 2nd CEO 3. 3rd CEO 4. 4th or later

Capital Japanese Yen

Q23) Your business field(s) (Check one or more) 1. 2. 3. 4.

Manufacturing Wholesale Retail Transport, communication

5. 6. 7. 8.

Construction Finance, insurance Real estate Corporate services

9. Individual services 10. Information services 11. Others ( )

Organizational Innovation in SMEs Generated by ICT

267

Q23–1) If you checked ‘1’ (manufacturing), what are the manufactured product(s)? 1. Food 2. Textile 3. Wool 4. Paper 5. Synthetic resin, rubber

6. Ceramic, rock and sand 7. Steel 8. Non-ferrous metals 9. Metals 10. Machinery and tools

11. Electric machinery and tools 12. Machinery for transport 13. Others ( )

Q24) What is the position of the current business relative to the initial (original) position when your company was founded? (Check one) 1. Same as the original businesses. 2. Has developed to the upstream (maker/vendor side) based on the original business. 3. Has developed to the downstream (consumer/user side) based on the original business. 4. Has developed horizontally (to different business) based on the original business. 5. Is a completely different business, and the original business has been abolished or reduced.

Q25) How many employees does your company have? Total

Regular/full-time

Part-time

Total number IT workers (ex. System administrators)

Q26) and Q27) are definitely important in this survey. Please make sure the following boxes are filled. Q26) Gross profit margin (gross margin/sales X 100) About

%

Return of sales (Operating income/sales X 100) About

%

268

Masatsugu Tsuji and Shoichi Miyahara

Q27) Please fill the boxes as much as your company policy allows. If you can’t give the numerical data in the upper boxes, please tell us the trend in the lower box. Sales

Gross margin Operating income

Equipment investment

IT investment

Export

Import

Last fiscal year (Three years before as 100) Trend 1. Increasing 1. Increasing 1. (Compared 2. Almost 2. Almost 2. to that of same same three years 3. Decreasing 3. Decreasing 3. before)

Increasing 1. Almost 2. same Decreasing 3.

Increasing 1. Almost 2. same Decreasing 3.

Increasing 1. Almost 2. same Decreasing 3.

Increasing 1. Increasing Almost 2. Almost same same Decreasing 3. Decreasing

That’s it. We really appreciate your cooperation We would be grateful if you would post this to us by 30 June. Please use the enclosed ‘return envelope’.

Notes 1. 2. 3. 4.

For AHP, refer to Saaty (1980, 1986), for example. Except Fujita et al. (1999). See Schumpeter (1934). This category was referred to as ‘orientation to data use’ or ‘data-using type’ in Tsuji et al. (2005) and Bunno et al. (2006a, 2006b). 5. −0.029 for ‘Unclear objectives of management’ and −0.034 for ‘ICT security is a major concern’. 6. The coefficient expresses the marginal contribution of an independent variable to the index, and its amount depends on the current level of the index. This is a one interpretation, but a more rigorous analysis is required to identify reasons of these problems for future research.

References Bunno, T., H. Idota, M. Tsuji, H. Miyoshi, M. Ogawa, and M. Nakanishi (2006a) ‘An Empirical Analysis of Indices and Factors of ICT Use by Small- and Medium-sized enterprises in Japan,’ Proceedings of 16th ITS Biennial Conference, Beijing, China, June 2006. Bunno, T., H. Idota, M. Ogawa, M. Tsuji, H. Miyoshi, and M. Nakanishi (2006b) ‘Index of the Diffusion of Information Technology among SMEs: An AHP Approach,’ The Proceedings of the 17th European Regional ITS Conference, Amsterdam, Holland, August 2006. Bunno, T., H. Idota, M. M. Tsuji, and M. Nakanishi (2007) ‘Factors and Policies for the Diffusion of Information and Communications Technology among Japanese SMEs,’ Proceedings of the 18th European Regional ITS Conference, Istanbul, Turkey, September 2007. Fujita, M., P. Krugman, and A. Venables (1999) The Special Economy: Cities, Region, and International Trade. Cambridge, MA: MIT Press.

Organizational Innovation in SMEs Generated by ICT

269

Japan Small and Medium Enterprise Management Consultants Association (2003) Report of Research on SCM Business Models for SMEs (in Japanese). Tokyo. Kuchiki, M. and M. Tsuji (eds) (2004), Industrial Clusters in Asia: Competition and Coordination. Basingstoke: Palgrave Macmillan. —— (2008) The Flowchart Approach to Industrial Cluster Policy. Basingstoke: Palgrave Macmillan. Saaty, T. L. (1980) The Analytic Hierarchy Process: Planning, Priority Setting, Resource Allocation. Newyork: McGraw-Hill. —— (1986) ‘Absolute and Relative Measurement with the AHP: The Most Livable Cities in the United States.’ Socio-Economic Planning Sciences, Vol. 20, No. 6, pp. 2–37. Schumpeter, J. A. (1934) The Theory of Economic Development. Oxford: Oxford University Press. Small and Medium Enterprise Agency (2004) ‘Project II for Promotion of ICT Use by SMEs’ (in Japanese). Tokyo, Ministry of Economy, Trade and Industry (METI). —— (2001, 2002, 2003, 2004) ‘White Paper on Small and Medium Enterprises in Japan’ (in Japanese), Tokyo, Ministry of Economy, Trade and Industry (METI). Tsuji, M., E. Giovannetti, and M. Kagami (eds) (2007) Industrial Agglomeration and New Technologies: A global perspective. Cheltenham: Edward Elgar. Tsuji, M., H. Miyoshi, T. Bunno, H. Idota, M. Ogawa, M. Nakanishi, E. Tsutsumi, and N. Smith (2005) ‘ICT Use by SMEs in Japan: A Comparative Study of Higashi-Osaka and Ohta Ward, Tokyo. OSIPP Discussion Paper, No. 06–05, Oskaka University.

8 The Role of the Specialized Markets in Upgrading Industrial Clusters in China Ke Ding

8.1

Introduction

This chapter attempts to provide some complementary cases to the Flowchart Approach. The typical pattern of the Flowchart Approach suggests that some machinery industrial clusters in East Asia appeared and were upgraded by introducing a foreign anchor firm into the industrial park (Kuchiki and Tsuji 2005, 2008). Then, endogenous R&D was accomplished by forging linkages with some research institutes, such as the universities. This typical pattern especially focused on the industrial park’s institutional conditions that support anchor firms and universities.1 The author agrees that in some industries (such as automobiles) and under certain initial conditions (such as a region without a tradition of manufacturing or trading), the role of the anchor firm is indispensable. In contrast to existing studies, however, we assert that the role of markets should not be ignored as well.2 In other words, the industrial clusters can be upgraded even within the fully competitive market, where the key actors are numerous small and medium-sized enterprises (SMEs). China provides a number of persuasive cases in this regard. Many important industrial clusters in this country have unique marketplaces referred to as specialized markets. They are generally wholesale markets, where each booth is highly specialized in the cluster’s local commodities.3 Its sellers and buyers are the countless number of SMEs. As Jin (2004) and Ding (2006a) suggested, in China’s largest specialized markets for daily necessities, apparel, leather goods, textile and metalworking, the industrial clusters were continuously upgraded and expanded with the development of these markets. Actually, almost all the clusters in developing countries have such marketplaces. However, because the institutions that support market transactions are not sufficiently provided, most of the marketplaces tend to be a large informal sector, hindering further development of the cluster. This is the 270

Specialized Markets in Upgrading Clusters in China

271

very reason why the markets were usually taken over by limited anchor firms. In comparison, how were the specialized markets in China able to overcome the failure of the market? What are the determining factors that allowed smooth operation of the market mechanism in China’s industrial clusters? This chapter attempts to answer these questions by analysing some typical cases. Section 8.2 introduces the features of specialized markets. Sections 8.3, 8.4 and 8.5 investigate three industrial clusters that have the largest specialized markets for each industry in China. All of these markets have undergone drastic changes that resulted in upgrading of the cluster. We state our conclusion in Section 8.6.

8.2

Features of the specialized market

As introduced in Section 8.1, the specialized market is a unique marketplace in many important industrial clusters in China. This section presents the features of the specialized market in detail, using the data on 68 markets within 53 typical industrial clusters in Zhejiang Province, where specialized markets and industrial clusters first appeared and are the most developed. First, the market structure of the specialized market is highly dispersed. As Table 8.1 indicates, in 1998, among the 68 markets in Zhejiang, at least 56 markets had no less than 100 booths. Of these, 14 markets had between 1,000 and 4,999 booths, and seven markets had no less than 5,000 booths. On the other hand, there are also a large number of buyers who make their purchases in specialized markets. Among Zhejiang’s above-mentioned 68 markets, according to limited data, there are two markets that have 100,000 buyers visiting per day.4 Additionally, there are five markets visited by 50,000, 15,000, 10,000, 8,000 and 50 merchants, respectively, per day.5 Their purchases are oriented toward the low-end demand. We can infer that no one company is able to control any others under this market structure. Secondly, the commodities of specialized markets were mainly distributed in the domestic market in 1998. As Table 8.2 clearly shows, among the 68 markets, 51 markets sold commodities to China’s domestic market,

Table 8.1 Number of booths in specialized markets in Zhejiang’s major industrial clusters (1998) Number of booths Number of markets

100–999 35

1,000–4,999 14

No less than 5,000 7

Source: Compiled by the author based on data from ZPMCEC (2000).

Unknown 12

272 Ke Ding

21 markets sold to developing countries, and 15 markets sold to developed countries. In addition to the domestic market, the specialized market seems to be superior at exploring the markets in developing countries. Thirdly, specialized markets developed simultaneously with the industrial cluster. Among Zhejiang’s 68 specialized markets, at least 36 markets have expanded or relocated its transaction buildings. Of these, 21 markets expanded or relocated these buildings multiple times. As the specialized markets expanded or relocated only when the business scale within the industrial clusters drastically expanded, we can confirm that specialized markets developed continuously with the industrial cluster. Active intervention from the local government is pointed out as the main characteristic feature of the specialized market. Among Zhejiang’s 68 markets, there are at least 38 markets where the local government is or has been in charge of construction and transaction management. Only five markets have been managed by firms from the beginning. Usually, the local government intervenes in a specialized market by establishing a managing committee composed of local government employees. Due to material constraints, the 1998 data presented above is limited. After that year, almost all the indexes are further advanced.6 Including other types of market, the average transaction volume of Zhejiang’s marketplaces increased from 69.5 million yuan to 179.0 million yuan during the period 1998 to 2005. At the same time, however, the number of marketplaces decreased from 4,619 to 4,008 as competition intensified (Jin 2007, p. 35). Based on the above features, the specialized market can be regarded as a microcosm of China’s emerging market – a highly expanding, fully competitive market. Its main players are numerous SMEs, aiming at the low-end demand. Transaction volume in the specialized market increased rapidly. New business relationships are constantly established. The local government plays an important role in the development of the specialized market. Hence, the study on the specialized market is essentially a study on the role of local government in a large emerging market. Table 8.2 Scope of specialized markets in Zhejiang’s major industrial clusters (1998, multiple)

Market scope Number of markets

Within the city

Beyond the city but within Zhejiang

Beyond Zhejiang but within the domestic market

Other developing countries

Developed countries

Unknown

5

1

45

21

15

15

Source: Same as Table 8.1.

Specialized Markets in Upgrading Clusters in China

273

8.3 Yiwu: Case study of a daily necessities cluster 8.3.1 Profile of Yiwu daily necessities cluster7 Yiwu is the world’s largest production and distribution centre for daily necessities. The Yiwu China Commodities City (Yiwu Market)8 has over 400,000 commodities in 1,901 categories from 43 industries. Its merchandise is distributed not only in China’s domestic market, but also in 212 countries and regions in the world (ZCCC Group 2006). Yiwu was a proper rural area at the end of the 1970s. However, after the local government formally established the Yiwu Market in 1982, Yiwu started its rapid cluster formation.9 During the period from 1982 to 1990, accompanying the increase in the number of booths in Yiwu Market from 700 to 8,900, 180 handcraft clusters, namely ‘One Village, One Product’, appeared in Yiwu. After the 1990s, the number of booths in the Yiwu Market increased to 58,000. Accordingly, the total transaction volume of the Yiwu Market reached 31.5 billion yuan. In this process, eight large industrial clusters were formed within Yiwu: socks, seamless underwear, suitcases and leather commodities, zippers, crafts, stationery and other daily necessities.10 The major companies in these clusters have since evolved into modern mass-production factories. The local government has played a key role in developing the Yiwu Market. They established several specific authorities to manage this market. Early on, the Yiwu government established a managing committee that consisted of several employees from local government departments. In 1994, they established a market-managing company called Zhejiang China Commodities City Group (ZCCC Group), which was listed on the Shanghai Stock Exchange in 2002. Meanwhile, local government departments such as the Administration for Industry and Commerce (AIC) and Quality and Technical Supervision (QTS) continued to maintain a link with this market. In order to support market transactions, both the managing committee and the managing company provided various infrastructures. The regular market was originally built in 1982. As business expanded, infrastructure constraints became obvious, and so a new, second-generation market went up in place of the original. Thus, future generations periodically replaced the earlier ones and the market is now in its sixth generation. On the other hand, the managing committee and managing company deregulated the private SMEs in the early 1980s. They also intervened in quality control, forgery exposure, etc. in the Yiwu Market. This chapter focuses on the activities of the managing committee in 1992 in particular. Despite the appearance of ‘One Village, One Product’, the tertiary sector of Yiwu continued to expand until 1992. After that, however, the secondary sector suddenly started growing. During the period from 1992 to 1998, the share of the secondary sector increased from 28.1 per cent

274 Ke Ding

to 50.7 per cent (ZUESRG 2008, p. 12). The year 1992 is undoubtedly the turning point of industrial cluster development in Yiwu. 8.3.2 Upgrading in classification of commodities11 Classification of commodities is a key factor in understanding the above structural change. Until the third-generation market, even though the Yiwu Market expanded several times, the classification of commodities was still very rough. In 1990, the market’s 8,000 booths were divided into only four industries: daily necessities, garments, knitwear and shoes (YYLGE Office 1992). Although all of the commodities were assigned specific spaces according to industry, it was easy to find the same type of commodity in any other commodity’s space. In some cases, when the booth-keeper was changed, the commodities of the booth would also change. Some of the smaller merchants who did not have a stable relationship with the manufacturers also changed their businesses often. As a result, the same type of commodities might be sold at different prices in different places of the market. The market mechanism obviously failed to function in this chaotic situation. Thus, when the fourth-generation market was planned for 1991, AIC, the major member of the managing committee, was determined to design an efficient way to classify by industry and location the large quantities and types of commodities sold at the Yiwu Market (Huahang Guishi). In the Yiwu Market, the property rights for all the booths belong to the Yiwu government. The booth-keepers only have the right to use the booths. Moreover, in China, landownership belongs to the public sector. Therefore, AIC was able to design a way to classify the commodities just as they wished. For Huahang Guishi, the staff of AIC visited numerous department and hardware stores to learn methods for classifying commodities by use, raw material and configuration, etc. They worked out a plan to classify the Yiwu Market into eight zones, where the commodities of 16 industries could be bought and sold. These industries were: (1) garments; (2) knitwear; (3) shoes; (4) socks; (5) ribbons; (6) wool yarn; (7) small hardware; (8) decorations; (9) daily necessities; (10) rainwear, bags and suitcases; (11) stationery and sporting goods; (12) cosmetics and other pharmaceutical goods; (13) buttons, zippers and other accessories; (14) toys; (15) lighters, watches and electronics; and (16) artificial flowers.12 Before moving into the new market, the booth-keepers were asked to register their business and obtain a license,13 as the first step in carrying out AIC’s plan. Initially, however, there was no reaction from the booth-keepers. Many of them were worried that their profits would drop due to the limit being placed on their scope of business. In order to solve this problem, AIC staff approached some of the leading merchants of the Yiwu Market, particularly those in the Communist Party. In the end, most of the boothkeepers were registered.

Specialized Markets in Upgrading Clusters in China

275

After the fourth-generation market opened in 1992, the AIC continued to take flexible measures in promoting the classification of commodities. At first, the newly built booths were allocated to more than 13,000 merchants by lottery. Then, they permitted all merchants, registered or not, to enter the new market. From the second month, they began permitting the resale or exchange of booth licenses. By the third month, all booth-keepers were required to show their license in the market. By taking the above measures, the situation in Yiwu Market underwent a dramatic change, resulting in the industrial cluster formation in Yiwu.14 First, the classification of commodities stimulated the booth-keepers to specialize in specific fields, so that their business would become more stable. As a result, during the period from 1992 to 1997, the share of booths having long-term business relationships with manufacturers increased from less than 30 per cent to 53.6 per cent. Meanwhile, at least 1,300 booths established their own factories (Ding 2006a, chapter 4). Secondly, by organizing areas of booths that dealt with the same commodities, the merchants were strongly motivated to develop newer and better commodities. For example, the number of shoe booths increased from 220 to 1,700 during the period from 1990 to 1992.15 It is reported that the booth-keepers of artificial flowers developed more than one type of new product almost every day. Thirdly, commodity classification enabled the expansion of the Yiwu Market. In 1994, the fourth-generation market was enlarged again. The new market was classified into 13 zones, where the commodities of 21 industries were bought and sold.16 Currently, as mentioned above, there are over 400,000 commodities in 1,901 categories from 43 industries. The number of booths increased to 58,000. With market expansion looking up, the booth-keepers accelerated their investment in the manufacturing sector. This classification of commodities has had a broad influence. In 1998, among Zhejiang’s above-mentioned 68 specialized markets, at least 18 markets had their commodities classified by industry and location. Of these, the transaction volume of five markets amounted to 100–1,000 million yuan. The transaction volume of 13 markets was no less than 1 billion yuan (ZPMCEC 2000). We can observe a clear correlation between the market transaction scale and the classification of commodities.

8.4

Danyang: Case study of an eyewear cluster

8.4.1 Profile of Danyang eyewear cluster17 Danyang is a county-level city, located in Zhenjiang City, Jiangsu Province. It is one of the largest eyewear industrial clusters in China.18 From the 1930s to the 1940s, a number of peasants living in the countryside of Danyang moved to Shanghai and Suzhou and worked at the eyewear factories as apprentices.

276 Ke Ding

At the beginning of the 1960s, some of these peasants returned to Danyang and began making lenses, frames and screws. After that, several eyewear factories appeared in Danyang. In 1977, there were five eyewear factories in Danyang. The production volume of these factories was 289,600 pairs of glasses and 387,600 pairs of frames. In 1985, the number of township eyewear factories amounted to 23. Their total production volume increased to 4,563,100 pairs of glasses and 2,592,500 pairs of frames, which accounted for one third of the total in China’s domestic market. The Danyang Eyewear Cluster continued its growth through the 1980s and 1990s. In 2004, the total production value of Danyang eyewear increased to 3 billion yuan, of which the export of eyewear amounted to more than 100 million dollars. The number of workers in the eyewear industry increased to 50,000. More than 1,000 factories and trading companies have appeared in this cluster. Of these, more than 400 are frame makers, more than 70 are CR-39 plastic lens makers,19 more than 100 are glass lens makers, more than 20 are screw makers, more than 20 are spectacle case makers, and 500 are trading companies and other supporting companies. It is publicized by the Danyang Optical Chamber of Commerce that the domestic share of Danyang plastic lenses has increased to more than 70 per cent. The world share of Danyang glass lenses and plastic lenses is 80 per cent and 50 per cent respectively (DOCC 2005, p. 112).20 The Danyang Optical Market (Danyang Market) is China’s largest optical market, and has played a crucial role in the development of Danyang’s eyewear cluster.21 In the 1970s, a small market for exchanging eyewear products was spontaneously formed around Danyang Station. The Danyang city government, the AIC, and the village near the station jointly established the formal Danyang Optical Market in 1982. This market opened in 1986 with a mere 35 booths. After several extensions, however, the total number of booths increased to more than 700 by 2003. In the same year, the transaction volume of this market reached 620 million yuan.22 Almost all eyewearrelated goods, including lenses, frames, parts and measurement instruments are sold in this market. Interestingly, at the first stage in 1987, only 25 per cent of eyewear-related commodities in the Danyang Market were produced in Danyang. However, by 2002, more than 200 booth-keepers had opened their own factories after starting out in this market. Many local manufacturers made use of this market as well. As a result, the local share of lenses and frames in the Danyang Market increased to 80 per cent and 70 per cent respectively in 2002.23 In order to understand the development of the Danyang cluster, it is necessary to clarify the factors that caused such a structural change. 8.4.2

Upgrading quality control24

The progress made in the quality control of eyewear is the most important factor in understanding the above structural change. Zhongguo Zhiliang

Specialized Markets in Upgrading Clusters in China

277

Wanlixing (ZZW) is a social movement that started in 1992. It was facilitated by China’s central government, major mass media, and famous companies, scholars and technologists.25 In 1995, several members of ZZW visited the Danyang Market and inspected the eyewear. Their report indicated that the examination pass rate for glasses in this market was, surprisingly, zero.26 CCTV and other major mass media outlets in China reported on this, and as a result, the Danyang cluster’s image suffered severe damage. Under great social pressure, the government of Danyang and the managing committee of Danyang Market had to take drastic measures to cope with this quality problem. In August 1996 they established a quality and technical supervision (QTS) department in Danyang Market.27 In September 1996, the QTS department and members involved with the Danyang Market28 inspected the quality of glasses of the market. They confirmed that the quality was indeed poor. The examination pass rate for lenses was only 45 per cent and for frames 60 per cent. The QTS department determined two reasons for the poor quality. The first was the poor quality control technology. It was estimated that 30 per cent of the inferior goods were poor for this reason. For example, few booths in the Danyang Market met China’s National Glass Lens Quality Standard of GB10180 at this time. Many booths did not even own the measuring instruments that are indispensable to the production and sale of lenses. In order to improve the poor QC technology, the QTS department decided to make annual visits to all of the factories related to the Danyang Market.29 The QTS staff usually checked the production system thoroughly, explaining the problems to the factory owners in detail. In the case of lenses, all booth-keepers who owned lens factories were required by the QTS to introduce electric lens measuring instruments.30 In 1996, the price of each instrument was 30,000 yuan. In cases where a booth-keeper was unable to afford the instrument, the QTS department in the Danyang Market or the Danyang QTS department, acting as a guarantor, would ask the sales agent to accept postponement of payment. Sometimes, they even helped booth-keepers in collecting debts from their customers.31 The second reason for the poor quality was a lack of QC awareness. The QTS department staff distinguished this as two types. The first type was due to low awareness of management. In concrete terms: (1) selling product A as product B in order to obtain a higher profit; and (2) processing the requested high-quality lenses, in spite of technological constraints, using poor-quality raw material. The second type was an extreme lack of awareness of trademark rights. There were only two trademarks in all of Danyang Market in 1996: Kangming and Huaguang. The first is the brand of a Singaporean company. The second is the brand of a manufacturer in Hubei Province, China. Under these circumstances, the QTS department took the following measures for improving QC awareness of the booth-keepers.

278

Ke Ding

First, the QTS department organized round-table talks with a number of booth-keepers. In specialized markets, merchants are usually connected by relationships based on geographical origin (Ding 2006a). Thus, the QTS department invited mainly the leaders of different regional merchant groups and tried to build a relationship of trust with those merchants. Afterwards, they explained to them the importance of QC. These talks lasted for two years. Secondly, the QTS department organized quarterly QC courses for all booth-keepers in the Danyang Market. Tuition and textbook fees were not charged. Thirdly, the QTS department advised all booth-keepers who owned factories to register their own trademarks. As a result, more than 400 booths had their own trademark by 2001. Some booths even had five or six trademarks. In addition to the above temporary measures, the QTS department regularly supervised and inspected the booths. In 2001, lenses were inspected every month, frames every six months, and the assembly of glasses every month.32 The booth-keepers had no advance notice of inspection, and all products were thoroughly examined. If quality problems were detected, the QTS department required the booth-keeper to improve the quality. If quality problems were detected a second time, a report was sent to the Danyang QTS department. The offender would be fined 10,000–100,000 yuan. In some cases, the booth-keeper was even forced to discontinue production.33 In the Danyang Market, not only the QTS department, but also other departments are in charge of the QC problem. For example, the AIC department places emphasis on exposing imitation products. Its staff members visit booths or other logistics points at random. If imitation goods are discovered, all the goods from that booth are confiscated and the offender is fined an amount ranging from 10,000 to 100,000 yuan.34 The members of the managing committee of the Danyang Market sometimes work jointly to handle the QC problem. Every year, a quality contest is held in the market under the sponsorship of the QTS department, AIC, Tax Office and the local village. The winner of this contest is given a certificate to be prominently displayed in their booth, which attracts consumers as it is a guarantee of quality. This provides the booth-keepers with a strong incentive to maintain QC. Because of the efforts of the members of the managing committee, the examination pass rate rose steadily. In 1998, the examination pass rate for lenses was 85 per cent, and 89 per cent for frames (JPZCCE Office 1999). In 2000, Danyang Market won the title of Jiangsu Shopping Rest Assured Specialized Market.35 In April 2001, the examination pass rate of the top 406 booths of Danyang Market was 95 per cent for lenses, 98 per cent for frames,36 and 59.8 per cent for reading glasses.37 The improvement in quality restored Danyang Market’s reputation. The number of new buyers visiting

Specialized Markets in Upgrading Clusters in China

279

this market has continuously increased, so that more and more eyewear merchants began investing in the manufacturing sector.

8.5 Yuyao: Case study of the mould industry 8.5.1 Profile of Yuyao mould cluster38 Yuyao is a county-level city located in Ningbo, Zhejiang Province. It is one of the largest plastic products and mould clusters in China.39 The origin of Yuyao’s plastic industry dates back to the 1960s when a few small Bakelite plastic factories appeared in the city. Accompanying the development of the plastic industry, the demand for moulds rapidly increased. As a result, a large number of factories specialized in mould production by the 1980s. According to the China Die and Mould Association, in the mid-1980s, the capacity of Yuyao mould production and injection accounted for one fourth of the total capacity in China. Since that time, Yuyao has been called the ‘Home of Plastic’ and the ‘Kingdom of Moulds’.40 In the beginning of the 1990s, in order to upgrade the Yuyao cluster, the China Light Industry Association (CLIA) invested more than 30 million yuan to build a model company named the Zhejiang Moulds Production Center (ZMPC). ZMPC received a complete suite of mould machines, with the capacity for integrated production. However, because the capacity was not fully utilized, and due to other managerial reasons, the company soon went bankrupt, leaving many skilled workers in Yuyao.41 In order to sustain daily life, most of these workers started their own businesses. At this time, however, due to poor linkage with other firms and intermediary social organizations, these SMEs were unable to immediately obtain raw materials, or place orders with outside suppliers for parts for the manufacturing process. Thus, they remained stagnant for a considerable length of time. In order to provide them with support to survive this situation, the Yuyao government established a marketplace specializing in moulds. The China Light Industrial (Yuyao) Moulds City opened in 1995, in cooperation with CLIA.42 The location of this market was an area where 150 of these SMEs had informally clustered. After that, the SMEs in the moulds industry seemed to develop very rapidly. The situation during the period from 1995 to 2000 is not clear. However, as indicated in Table 8.3, during the period from 2001 to 2006, the number of mould companies in Yuyao increased from about 1,000 to more than 1,300, and the number of workers increased from about 20,000 to more than 45,000. At the same time, mould production volume increased from 800 to 5,200 million yuan. By comparing the number of mould companies in the Yuyao cluster and in the Yuyao Market, we can easily observe the correlation between the Yuyao Market and the Yuyao mould cluster. As Table 8.3 indicates, the number of newly started businesses in the Yuyao Market is larger than

280 Ke Ding Table 8.3

Year 2001 2002 2005 2006

Profile of the Yuyao Moulds cluster

Production volume in Yuyao cluster (million yuan) 800 1,530 3,000 5,200

Number of workers in Yuyao cluster More than 20,000 More than 30,000 More than 50,000 45,000

Number of companies in Yuyao cluster

Number of companies in Yuyao Market

More than 1,000 More than 1,200 More than 1,300 More than 1,300

220 521 658 More than 700

Sources: 2001: CPCIC (accessed 17 January 2007); 2002: ZQMW (accessed 17 January 2007). 2005: Gong and Wan (accessed 17 January 2007); 2006: ‘Yuyao Mojucheng Shixian “Yizhanshi” Fuwu Zhizao’ [The Yuyao Mould Market Provides ‘One Stop Service’ to the Manufacturing Sector], www.chemhello.com (accessed 28 January 2008).

that in the Yuyao cluster. The number of companies in the Yuyao Market increased more rapidly. This means that besides the fact that most of the newly started companies are located in the Yuyao Market, many outside companies have moved into this market too. In other words, the Yuyao Market has successfully taken the place of ZMPC in upgrading the Yuyao cluster. 8.5.2 Upgrading in mould value chains The Yuyao Market is managed by a committee whose members are staff from the Yuyao government. By analysing the activities of this committee, we can understand in detail how the Yuyao Market upgraded the Yuyao cluster.43 In concrete terms, this committee has taken the following measures: First, the managing committee built two raw material submarkets in the 1990s, having a gross area of 46,000 m2. As of 2007, they had invited more than 80 domestic and overseas raw material producers to set up sales outlets within these markets, including famous companies like Bao Steel from Shanghai and Assab Steel from Singapore. As Table 8.4 indicates, during the period from 2001 to 2005 the number of raw material companies nearly doubled. The transaction volume and the weight of raw materials increased correspondingly. It is well known that the price for China’s metal raw materials drastically increased during this period. The price of steel in China has risen 3–4 times during the same period. However, because the clustering of companies offering the same products caused intense competition, the price of raw materials in the Yuyao Market has risen no further than 1.1 times. Secondly, the managing committee built a mould technology and machine exhibition centre in 1999. It covers 4,600 m2. In 2006, more than 100 types of mould from Yuyao’s top 12 companies and those of other SMEs were exhibited at this centre.44 Since 1999, the centre has also held a yearly trade fair for moulds.45

Specialized Markets in Upgrading Clusters in China

281

Table 8.4 Raw material businesses in the Yuyao Market

Year

Number of raw material companies

2001 2005

More than 40 More than 80

Transaction volume of raw material (million yuan)

Weight of raw material (1000 t)

More than 500 More than 900

Less than 60 100

Sources: 2001: CPCIC (accessed 17 January 2007); 2005: Gong and Wan (accessed 17 January 2007).

Table 8.5 Division of labour in the Yuyao Market

Year

Total number of companies in the Yuyao Market

2001 2005

220 658

Number of mould production companies 50 More than 100

Number of mould processing companies 100 More than 300

Number of companies producing raw material and others 70 Less than 258

Sources: Total number of companies in the Yuyao Market: Same as Table 8.3. The other data of 2001: Shao and Ding (accessed 17 January 2006). The other data of 2005: Interview with the vice president of the Moulds Association of Yuyao in July 2006.

Thirdly, the managing committee established an information centre in 2003, in cooperation with the Zhejiang Province Science Department. It also established its own website. By 2006, the centre had accepted 133,200 members and announced more than 300,000 pieces of information.46 The high accessibility of raw materials and information resulted in a deeper division of labour in the Yuyao Market. As Table 8.5 indicates, during the period from 2001 to 2005, the number of mould production companies more than doubled. At the same time, the number of mould processing companies increased by more than a factor of three. The number of companies producing raw materials and other types increased as well. By 2005, the value chain of moulds in Yuyao Market evolved into design, software development, wire cutting, NC line, tools, parts, and so on. Each manufacturing process is taken care of by specialized companies. The deepening of the division of labour directly resulted in lower production costs. Mould prices in Yuyao were only one-third of that in Japan and half of that in Guangdong Province in 2002. The difference in estimated prices between the companies in Yuyao Market in this year is not outside the scope of 10–15 per cent.47 In recent years, the managing committee noticed their increasing role in organizing the mould value chains. In 2006, they started a project to establish the Yuyao Market as an innovation platform consisting of ‘five centres

282

Ke Ding

and one base’. The market was simultaneously enlarged in line with this project. In addition to the above-mentioned exhibition centre and information centre, they established the following institutions. First, the managing committee built a 5,000 m2 precision processing base, at a cost of 50 million yuan. They encouraged the top companies in Yuyao to move their unused machines to this zone so that the local SMEs could share the excess capacity.48 On the other hand, students of local colleges could also make use of this precision processing zone as their training centre. By November 2007, 50 sets of machines had been introduced to this zone. Secondly, the managing committee established a training centre for skilled workers, in cooperation with the Baotou Technology College and the Yuyao Education Department. By November 2007, this centre had trained more than 500 mould workers. It has also introduced 45 students who learned mould technology such as CAD/CAM in college.49 Thirdly, the managing committee established an inspection and measurement centre, so that they would be able to provide authoritative reports to solve technological related problems. This centre is working in cooperation with the Weapons Science Academy, Ningbo Branch for analysing the ingredients of metal materials; and the Yuyao QTS centre for measuring the length and cubic content of moulds. Fourthly, the managing committee established an R&D centre for mould innovation. Currently, it is working in cooperation with the Automobile Institute of Zhejiang University for the recommendation of new technologies; the Beijing Machinery Institute for the development of new software; and the National Level Laboratory of Moulds of the East China University of Science for researching the fundamental theory of moulds. As stated above, the purpose of the managing committee is to develop a platform that drives chain governance within the Yuyao mould cluster, exactly the way a lead firm acts in a global value chain. The only difference is that the former accomplishes it through the market mechanism, and the latter relies on the power of big business.

8.6 Conclusions In developing countries, there is usually a big discontinuity in the process of cluster development. According to the author’s understanding, the core idea of the Flowchart Approach is just to provide a platform that overcomes various difficulties, in order to urge a cluster from an immature step to the next step onward. Due to different initial conditions, there could be various forms of such a cluster support platform. The typical pattern of the Flowchart Approach shows that when an industrial cluster is dominated by an anchor firm, a well-designed industrial park can work as such a platform. In addition to this typical pattern, the cases of the specialized markets indicated other

Specialized Markets in Upgrading Clusters in China

283

possibilities. Namely, when a cluster has numerous SMEs as its key actors, a platform that enables the market mechanism to work smoothly would be more desirable. Sharing the core idea, this specialized market dynamism can largely enrich the framework of the Flowchart Approach. A possible theory to explain the specialized market is the Market Design approach. McMillan (2002), as the best work in this field, pointed out that for a market to function well, it is necessary to design a platform that has five elements. Namely, that information flows smoothly; property rights are protected; people can be trusted to live up to their promises; side effects on third parties are curtailed; and competition is fostered. This approach also emphasizes that help from the government is essential if the market is to reach its full potential. As stated, ‘To reach a degree of sophistication, its (market’s) procedures need to be clarified and an authority given the power to enforce them. Only when the informal rules are supplemented by some formal rules can a market reach its full potential, with transactions being conducted efficiently and complex dealings being feasible’ (McMillan 2002, p. 13). The role of local government in the above-mentioned three cases can be well explained by this theory. In the case of Yiwu, the AIC worked out an effective way to classify the same type of commodities into the same area. As a result, the buyers can easily obtain information on daily necessities. The booth-keepers can also easily obtain their competitors’ information. Further competition and new ideas are thus stimulated. In the case of Danyang, the QTS and other departments tried every means available to maintain the credibility of eyewear transactions in the market. They provided the information on quality control to the booth-keepers. They also made available the information on the quality of eyewear to the buyers. Trademark registration was encouraged and rules were enacted in terms of fines imposed on offenders of the QC standard. As a result, the reputation of Danyang eyewear improved dramatically. In the case of Yuyao, the local government clearly recognized the Yuyao Market as a market-based platform in supporting SMEs. By forming linkages with various institutions inside and outside Yuyao, related firms from the upstream to the downstream of the moulds industry have been condensed into the Yuyao Market. New markets for labour, equipment and information have also been established. Thus, a more complete and sophisticated value chain of moulds was gradually formed. In the above cases, what the local governments did eventually met one or more of the above five elements. The Market Design approach, in this sense, is quite an appropriate theory in explaining the specialized market. However, we still instinctively noticed that there are considerable differences in the market mechanism between the specialized market cases and those of the Silicon Valley (Saxenian 1994; Yonekura 1999) or Third Italy (Piore and Sabel 1993).

284

Ke Ding

This could be concluded as the difference between a large emerging economy and advanced countries. Lastly, we attempt to make some tentative explanations for this issue and discuss how other developing countries can learn from China in upgrading their clusters through the market mechanism. First, the demand conditions are quite different between China and advanced countries. The market potential of advanced countries is considerably uncharted. Firms thus need to take more risk and encounter more failures. In contrast, the size of China’s domestic market is constantly expanding. However, its majority share is still the low-end demand. Thus, once the infrastructures and institutions for the specialized market are improved, not only the scale but also the content of transactions advances dramatically. To fully exploit the potential of domestic demand in the developing countries, it is necessary to clarify how China formed a strong linkage between its huge domestic market and these specialized markets.50 Secondly, the market support platform is usually invisible and wideranging in advanced countries (such as e-commerce websites or business associations). In comparison, such a platform developed as a concrete marketplace in China. As mentioned in each case, the local government enlarged and improved the transaction space of the specialized markets again and again by taking advantage of the flexibility of landownership and the ownership of booths in China. To learn from China’s experience, we must further investigate the property rights issue of real estate in the industrial clusters. Thirdly, the local governments in China have a strong incentive to intervene in economic development. They have more resources and information than firms do, and thus are more powerful, but they also have as little negative impact as possible on the autonomy of private firms. Contrary to this, in advanced countries, the firms are more active and the role of the government is limited in providing laws and other formal rules. In order to fully utilize the potential of the public sector, the features of the activities of China’s local government must be elucidated.

Notes A draft of this chapter has been presented as an IDE Discussion Paper, No. 88, Chiba: IDE-JETRO. 1. The role of big business seems to be emphasized by most of the important industrial cluster study groups that focus on the developing countries. For example, the Global Value Chain groups show that an industrial cluster can acquire development opportunities by trading with foreign lead firms. They scrutinized the various relationships between these two parties and classified them into patterns. (See the Global Value Chain Initiative website: http://www.globalvaluechains.org, accessed 2 October 2006.) On the other hand, the cluster-based industrial development group believes that clusters in developing countries can foster their own big business as well (Sonobe and Otsuka 2004). They empirically proved that after

Specialized Markets in Upgrading Clusters in China

2.

3. 4. 5. 6. 7. 8. 9. 10. 11.

12. 13. 14. 15. 16. 17.

18. 19. 20.

21.

22.

285

keen price competition in the quantity expansion phase, some large enterprises would appear and lead the whole cluster into the quality improvement phase. This chapter defines market transaction as a voluntary exchange: each party can veto it, and (subject to the rules of the marketplace) each freely agrees to the terms, in accordance with McMillan (2002, p. 6). In this definition, decision-making autonomy is the key in judging whether a transaction is based on the market or not, and its extent in the market. According to this definition, when an industrial cluster is governed by limited anchor firms, the decision-making autonomy of the SMEs in the same cluster must be constrained to some extent. Thus, the major transactions within this cluster are not the exact market transactions. Recall the various European markets introduced in Fernand Braudel’s books. This is the maximum number. This is the maximum number. The share of exports to both developed and developing countries in the specialized market has increased. For details of the development of the Yiwu Market, see Lu et al. (2003) and Ding (2006a, 2006b). Yiwu Market is one of Zhejiang’s above-mentioned 68 markets. As mentioned below, it can be regarded as a clustering of several specialized markets. In the 1970s, some vendors spontaneously formed two periodic markets in Yiwu. The formal Yiwu Market was converted from one of these two markets. From my interview with an officer of the Yiwu Economic Development Department in September 2007. This part is mainly based on Xinhuanet (2006) and on my interview in September 2007 with Mr He Zhangxing, who was in charge of the AIC in 1992. In the following paragraphs on the Yiwu Market, sources are only noted when being cited from other sources. The data of the classification plan in 1992 is cited from Zhang et al. (1993). Mr He did not explain the registration in detail. It is assumed that the business scope of booths was not strictly fixed until this time. The following three points were mainly pointed out by Mr He Zhangxing and then were reorganized and strengthened for this chapter. The number of shoe booths in 1990 is cited from the YYLGE Office (1992). 1994 data is cited from the ZPZHD Committee (1997). The profile of the Danyang Cluster is mainly based on Xu and Xu (2005). I partially joined in their fieldwork. I have also done my own fieldwork in Danyang since 2001. In the following paragraphs on the profile of the Danyang Cluster, sources are only noted when being cited from other authors. Other typical eyewear clusters are the Duqiao Cluster and the Wenzhou Cluster in Zhejiang province, and the Shenzhen Cluster in Guangdong province. The production volume of plastic lenses amounted to 150 million pairs. The data for the production volume of plastic lenses is also cited from this source. The credibility of these data still needs to be checked against exact worldwide data. Danyang Cluster is located in the southern part of Jiangsu province. This area’s pattern of industrialization is similar to that of the northeast part of Zhejiang (Ding 2006a, chapter 1). Thus, the basic features of the specialized market in Zhejiang derived from the above 68 markets can be applied to this market as well. As mentioned below, a large number of booths in this market have their own factories. After the first transaction, most buyers trade directly with the factories,

286

23. 24.

25.

26.

27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

38.

39. 40. 41. 42.

43.

Ke Ding skipping the market. Thus, the share of the transaction volume of the Danyang Market is not as large in the Danyang cluster. This paragraph is based on my interview with a booth-keeper in the Danyang Market in May 2002. This part is mainly based on my interview with the person in charge of the QTS department of the Danyang Market in April 2001. In the following paragraphs on QC in the Danyang Market, sources are only noted when being cited from other authors. This movement placed emphasis on the exposure of imitation goods and goods of poor quality. In order to obtain the true information, members of ZZW would visit a factory, market, or department store without advance notice, and thoroughly inspect the quality of commodities in these places. The results were reported to China’s major mass media outlets. According to the interviewee, most of the booths in this market closed their shops on this day. This data was based on the inspection results of merely three booths. All the staff of this department came from the QTS department of Danyang city. Including public security, the Office for Industry and Commerce, the Tax Office, and the local media. Of course, their activity was limited to Danyang. Most of the instruments were made by the Japanese manufacturer TOPCON. The person in charge of the QTS department in the Danyang Market, whom we interviewed, had helped 22 booth-keepers by 2001. However, the parts are not inspected. This information on fines is based on an interview with a booth-keeper in the Danyang Market in April 2001. This information on the AIC is based on an interview with a booth-keeper in the Danyang Market in April 2001. Only four markets won this title in Jiangsu Province. The remaining 2 per cent are imitation goods. The rest of the booths’ glasses (the majority are sunglasses, reading glasses and parts) were usually purchased from other eyewear clusters such as Wenzhou. The pass rate of these booths was 60 per cent, which is still low. This part is mainly based on my interview with the vice manager of the managing committee of the Yuyao Market and the vice president of the Mould Association of Yuyao in July 2006. In the following paragraphs on the profile of Yuyao’s mould industry, sources are only noted when being cited from other authors. Huangyan Cluster in Taizhou, Zhejiang is also famous for its moulds industry. The information on China Die and Mould Association is cited from CPCIC (accessed 17 January 2007). The information on ZMPC is partially cited from Shao and Ding (2002, accessed 17 January 2007). Yuyao Market is one of Zhejiang’s above-mentioned 68 markets. As buyers usually go to this market and place their orders directly, the area is treated as a specialized market in this chapter. Meanwhile, as most booth-keepers produce moulds inside this market, the Yuyao Market is sometimes referred to as an industrial park. This part is mainly based on my interview with the vice manager of the managing committee and the vice president of the Mould Association of Yuyao in July 2006. In the following paragraphs on the Yuyao Market, sources are only noted when being cited from other authors.

Specialized Markets in Upgrading Clusters in China

287

44. The information on the usage of the exhibition centre is cited from Gong and Wan (accessed 17 January 2007). 45. The information on the trade fair is cited from ZQMW (accessed 17 January 2007). 46. The data on the number of members and pieces of information is cited from Gong and Wan (accessed 17 January 2007). 47. The information on mould prices in Yuyao is cited from China Moulds Net (accessed 18 January 2007). 48. If there is only one machine in Yuyao, the zone would be used without requiring a fee, and the owner would receive a subsidy. If there is no less than two of the same type of machine in Yuyao, the owner would be required to pay a small fee for using the space. Of course, the SMEs must pay a rental fee. 49. The details about the precision processing base and the training centre are cited from ‘Yuyao Mojucheng Shixian “Yizhanshi” Fuwu Zhizao’ [The Yuyao Mould Market Provides ‘One Stop Service’ to the manufacturing sector] (www.chemhello.com, accessed 28 January 2008). 50. As for this issue, Ding (2006b) provided a thorough case study on the domestic distribution system of the Yiwu cluster.

References China Moulds Net ‘Yuyao Mojucheng Chenggong Mijue’ [The Key to Success of Yuyao Moulds City], http://www.tyzb.com.cn/info8710.htm (accessed 18 January 2007). China Plastic City Information Center (CPCIC) ‘Zhongguo Qinggong Momu Cheng – Mojucheng Jianjie’ [Brief Introduction to China Light Industrial (Yuyao) Moulds City], http://cpe.21cp.net/zgqgmjc/jianjie.htm (accessed 17 January 2007). Danyang Optical Chamber of Commerce (DOCC) (2005) Danyang Glasses – China. Danyang: Docc. Ding, Ke (2006a) Gendai Chuugoku niokeru Sanchikeisei Bunseki no Tame no Ichishiron [A Study on the Formation of Industrial Clusters in Modern China]. Nagoya University, Doctoral thesis. —— (2006b) ‘Distribution System of China’s Industrial Clusters: Case Study of Yiwu China Commodity City.’ IDE Discussion Papers, No. 75, Chiba: IDE-JETRO. Gong, Ning and Keda Wan (2006) ‘Suliao Wanguo Zhanling Mojuye Zhigaodian’ [Yuyao: A ‘Plastic Kingdom’ that Occupied a Commanding Height of the Moulds Industry], http://unn.people.com.cn/GB/14748/4955645.html (Original source is East China News, accessed 17 January, 2007). Jiangsu Province Zhenjiang City Chronicles Editing Office (JPZCCE Office) (1999) Zhengjiang Nianjian 1999 [Zhengjiang Yearbook 1999]. Beijing: Fangzhi Press. Jin, Xiangrong (2004) ‘Sekkoushou niokeru Sengyouka Sangyouku’ [The Industrial Districts in Zhejiang Province], in Chuugoku Kougyouka no Nosonteki Kiso: Choukou Karyuuiki wo Chuxin ni [Rural Basis of Chinese Industrialization: With Particular Reference to Downstream Yangtze River Areas], Takeuchi Johzen Ed. Nagoya University East Asian Study Series I. Nagoya: Nagoya University East Asian Industrialization Research Project, pp. 9–37. —— (2007) ‘Zhejiangsheng de Chanyejiqun – Yingdui Chanyeshengji Tiaozhan de Zhongxiaoqiye’ [The Industrial Clusters in Zhejiang – Challenge to Industrial Upgrading and Local SMEs], in Dangqian Zhongguo Chanyeshengji Qushi Yanjiu [A Study on the Trend of Industrial Upgrading in China], ed. The Research Project on

288

Ke Ding

Chinese Enterprises: The Quest for Industrial Upgrading amid Transition, Chiba: IDE-JETRO, JRP Series, pp. 31–104. Kuchiki, Akifumi and Masatsugu Tsuji (eds) (2005) Industrial Clusters in Asia: Analyses of Their Competition and Cooperation. London: Palgrave-Macmillan. —— (eds) (2008) Flowchart Approach to Industrial Cluster Policy. London: PalgraveMacmillan. Lu, Lijun, Xiaohu, Bai and Zuqiang Wang (2003) Shichang Yiwu – Cong Jimaohuantang Dao Guoji Shangmao [Market Yiwu – From Jimaohuantang (Exchange of the Feathers of Roosters with Sugar) to International Business]. Hangzhou: Zhejiang Peoples Press. McMillan, John (2002) Reinventing the Bazzaar. New York: W.W. Norton & Company. Piore, Michel J. and Sabel, Charles F. (trans. Yamanouchi Yasushi, Ishida Atsumi, and Nagai Koichi) (1993) The Second Industrial Divide: Possibilities for Prosperity. Tokyo: Chikuma Press. Saxenian, Annalee (1994) Regional Advantage. Cambridge: Harvard University. Shao, Jie and Zhiming Ding (2007) ‘Yuyao Moju Cheng de Chengben Youshi Laizi Nali?’ [Where Did Yuyao Moulds City Gain Its Cost Advantage?], http://www.zjol.com. cn/gb/node2/node43163/node44855/node74692/node74697/userobject15ai747904. html (accessed 17 January, 2006). Sonobe, Tetsushi and Keijiro Otsuka (2004) Sangyou Hatten no Rutu to Senryaku – Nicchutai no Keiken ni Manabu [Roots and Strategies of Industrial Development – Lessons from the East Asian Experience]. Tokyo: Chiizumi Press. Xinhuanet Jiangsu Channel (2006) ‘Siwei Gongchen Zonglun Yiwu Shichang De QianshiJinsheng (Xia)’ [Four Persons of Merit Talk About the History of the Yiwu Market (II)], http://www.zj.xinhuanet.com/tail/2006–06/07/content_7202907.htm (accessed 16 January 2007). Xu, Yuanming and Zhiming Xu (2005) ‘Zhonguo Jiangsusheng Nabu Diqu Chanyejiju Diaochabaogao’ [A Research Report on the Industrial Clusters in the South Part of Jiangsu Province, China], A Joint Study Report Submitted to Nagoya University, East Asian study project (unpublished). Yiwu Yearbook Leading Group of Editing Office (YYLGE Office) (ed.) (1992) Yiwu Nianjian (1986–1990) [Yiwu Yearbook (1986–1990)], pp. 257, 261. Yonekura, Seiichiro (1999) Keiei Kakumei no Kouzou [The Structure of the Managerial Revolution]. Tokyo: Iwanami Shinsho. ‘Yuyao Mojucheng Shixian “Yizhanshi” Fuwu Zhizao’ [The Yuyao Mould Market Provides ‘One Stop Service’ to the manufacturing sector], http://www.chemhello. com (accessed 28 January 2008). Zhang Wenxue. et al. (ed.) (1993) Yiwu Xiaoshangpin Shichang Yanjiu – Shehuizhuyi Shichangjingji zai Yiwu de Shijian [A Study on Yiwu Commodity Market – The Practice of the Socialist Market Economy in Yiwu]. Beijing: Qunyan Press, pp. 267–269. Zhejiang China Commodity City Group Co., Ltd. (ZCCC Group) (ed.) (2006) 2006 Nian Zhongguo Xiaoshangpincheng Shangpin Mulu [Commodities Catalogue of China Commodity City in 2006]. Yiwu: ZCCC Group. Zhejiang Province Market Chronicle Editing Committee (ZPMCEC) (ed.) (2000) Zhejiangsheng Shichang Zhi [Zhejiang Province Market Chronicle]. Beijing: Chronicle Press. Zhejiang Province Zhengxie Historical Data Committee (ZPZHD Committee) (ed.) (1997) Xiaoshangpin, Dashichang – Yiwu Zhongguo Xiaoshangpincheng Chuangyezhe Huiyi [Small Commodities, Big Market – the Memoirs of the Founders of Yiwu China Commodity City]. Hangzhou: Zhejiang Peoples Press, p. 97.

Specialized Markets in Upgrading Clusters in China

289

Zhejiang University Economics School Research Group (ZUESRG) (2008) ‘Yiwu Huangyan Diqu de Chanyejiqun: Yingdui Chanyeshengji Tiaozhan de Zhongxiaoqiye’ [The Industrial Clusters in Yiwu and Huangyan: Challenge to Industrial Upgrading and Local SMEs]. In the Research Group on the Reform and Upgrading of Chinese Enterprises, ed. Dangqian Zhongguo Chanye Shengji Qushi Fenxi: Hangye Anli Yanjiu (ii) [An Analysis on the Trend of Industrial Upgrading in Current China: Case Study of Industries (ii)]. Chiba: Institute of Developing Economies, Japan External Trade Organization (IDE-JETRO), Joint Research Program Series 144. Zhongguo Qinggong Moju Wang (ZQMW) ‘Zhongguo qinggong momu Cheng’ [China Light Industrial (Yuyao) Moulds City], http://www.mouldscity.net/city/ j1.htm (accessed 17 January 2007).

9 Industrial Clusters and Workplace Training to Expand Innovation Capability: Evidence from Manufacturing in the Greater Bangkok, Thailand Tomohiro Machikita

9.1

Introduction

This chapter studies the inside of regional capacity building in terms of human resources. The Flowchart Approach to industrial cluster policy has found that capacity building is a key factor to the success of the industrial cluster policy. Especially, development of human resources has been one of the prime essentials for a regional growth strategy. A great deal of effort has been made on capacity building in terms of human resources for regional as well as national economies. What we need next is firm-level analysis to study the inside of regional capacity building. Although Part I of this book also studies the learning from university–industry linkages, little is known about another pathway of capacity building in terms of firm-provided training. Recent years have seen a renewal of interest in firm-provided training as a source of expanding innovation capability. Firm-provided training plays a fundamental role of stimulating the endogenous technological changes and adoption of new technologies. What seems to be lacking, however, is studying the effect of economies of agglomeration on training. This chapter estimates the effects of economies of agglomeration on firmprovided training using evidence from industrial clusters in the greater Bangkok area, in Thailand. Industrial clusters in the greater Bangkok area have a unique aspect in terms of local labour market conditions. I find there are different types of industrial clusters with high and low turnover rates. I test whether different types of industrial clusters in terms of turnover make different evidence of training or not. This test suggests different implications of the success of the industrial cluster policy as follows. Clustering 290

Clusters and Training to Expand Innovation Capability

291

firms and industries are recognized as the driving force behind regional growth. Furthermore, labour pooling also plays a significant role in regional growth, by improving matching quality. Clustering firms and industries also stimulate the spread of information on the quality of goods and labour. Trading goods and employee mobility within an industrial cluster would cause knowledge and ideas to spill over among firms and workers. By contrast, clustering is not always a panacea for regional growth. If labour poaching reduces the optimal investment in training and these result in under-investment in human capital, then the clustering of firms does not lead to regional growth, based on upgrading incumbent workers’ skills. This is the central problem of policy-making. Therefore the Flowchart Approach to industrial cluster policy needs to test whether there are fewer training evidences with high turnover rates due to labour poaching within industrial clusters. We shall now look more carefully into the background of the research and policy. Markusen (1996) reviews the relationship between local labour market conditions and several types of industrial districts: Marshallian industrial districts; Hub-and-spoke districts; and Satellite industrial platforms. Markusen (1996) hypothesizes that there are different labour turnover rates by types of industrial districts due to the difference in labour market flexibility inside the districts. Markusen (1996) does not explicitly review the evidence of innovation, the adoption of new technologies, or workplace training because she focuses only on the relationship between labour mobility and agglomeration economies. The labour-based theory of agglomeration requires the accumulation of rigorous empirical evidence of training. Brunello and Gambarotto (2007) raise a similar question and tested the negative effect of spatial agglomeration on employer-provided training, using UK data. The evidence from the UK suggests that training is less frequent in economically denser areas. Brunello and de Paola (2008) also present a search-match training model and test whether labour turnover in denser areas reduces training, using a dataset from Italian firms. The evidence from Italy suggests that there is less training in provinces with a higher labour market density, measured by the number of employees per square kilometre. Interestingly, Brunello and de Paola (2008) raise the possibility of symbiosis between local knowledge spillover, based on an agglomeration economy and training. This generates a positive correlation between the incidence of training and local density. However, this effect is smaller than poaching externality that reduces firm-provided training. The purpose of this chapter is to examine the effect of thick labour market on the incidence of firm-provided training as innovation capacity building. To do this, I verify that the incidence of training is not frequent in industrial clusters with higher turnover rates. I also verify that the returns to training are lower in industrial clusters with higher turnover rates. It is useful for us to carefully revise cluster policy to study on firm-provided

292

Tomohiro Machikita

training in different level of market thickness. The Flowchart Approach to industrial cluster by Kuchiki and Tsuji (2005, 2008) emphasize not only the need for industrial agglomeration but also university–industry linkages to foster innovation. But the matter is not quite as simple as this approach suggests. The Flowchart Approach to the formation of industrial cluster lacks the analysis of negative side effect of labour market thickness. Agglomeration of industry and workers stimulate workers to do labour turnover more. This also stimulates firms to train workers less. To implement cluster-based capacity building policy, understanding the correlation of agglomeration and training is required. This chapter is a necessary complement to the Flowchart Approach in terms of capacity building within clusters.1 The results of this chapter can be summarized as follows. The empirical method is simple. I regress the on-the-job or off-the-job training incidences on industrial cluster dummy variables and individual characteristics, using the probit model. Compared to the food industry, the auto-parts industry provides more on-the-job and off-the-job training. On the other hand, the PC and HDD industries provide less on-the-job and off-the-job training. Because of the employment of under 40-year-old workers, the average tenure in the PC and HDD industries is typically shorter than in the food and auto-parts industries. The average tenure in each industry is a proxy of labour turnover. There are correlations between industries with higher labour turnover rates and little evidence of firm-provided training. I conclude that this result is derived from the effects of labour poaching. The return to training also plays a crucial role in determining the incidence of on-the-job and off-the-job training. If the return to training is insufficient to compensate for the training costs of skilled workers, then the firm gives up training and switches to labour poaching from the industrial cluster. Next, based on wage regression, I find evidence that the return to training is lower in industries with high turnover rates. There are the two possible ways to explain the lack of evidence of firm-provided training in the PC and HDD industries: (1) the high rate of labour turnover reduces the incentive to invest in human capital; or (2) the lower rate of the return to training also reduces the incentive to invest in human capital and increases the incentive to poach labour outside the firm. The latter effect is useful in explaining the evidence of the lower rate of return to training in an industry with a high turnover rate. Poaching and turnover behaviour reduces the tenure of employees. However, I cannot identify the two primary causal effects of labour turnover and return to training from the incidence of training. If the assumption of industry-specific human capital for workers is valid, workers tend to switch firms within the same industry. Both the thick labour market and industry-specific human capital stimulate employee mobility within the same industry and between firms in the greater Bangkok area. The contribution of this chapter is that I characterize a typology of workplace training in industrial clusters. The case study of Bangkok metropolis

Clusters and Training to Expand Innovation Capability

293

provides clear evidence in understanding why there are differences in the incidence of training between different types of industrial clusters. This makes a key contribution to literature of the training and skill development in economic geography. This chapter is organized as follows: Section 9.2 presents related literature; Section 9.3 describes a simple model to derive testable hypotheses of labour market thickness and workplace training; Section 9.4 shows the structure of the dataset; Section 9.5 presents empirical results of the training incidence and the returns to training; In the final section, we conclude the chapter and discuss any remaining issues. Further policy issues based on the findings are also discussed.

9.2

Related literature

This chapter has relationships with two research lines: (1) labour theorybased agglomeration; and (2) the Flowchart Approach to the formation and upgrade industrial cluster that emphasizes the role of university and business linkages within agglomeration. In the last part of this section, I will show the importance of studying with labour theory-based agglomeration to develop industrial cluster policy. Agglomeration of industry and workers cause positive and negative externality to the labour market. A considerable number of studies have been made on pecuniary externality between physical capital and human capital investment. Technological externality of knowledge spillovers within cluster is also empirically examined. These are the positive side of agglomeration economies. It is widely known that thick market externality affects firm-provided training in each cluster, that is, poaching externality. This is the negative side of agglomeration. Although cause and consequences of clustering have been important to design cluster policy, there is not substantial literature to evaluate the negative side effects of clustering firms on human capital accumulation. Literature on workplace training across clusters is also not common in developing countries, although numerous studies have been made in European countries. The empirical reason for this is that it is difficult to collect appropriate datasets on firm- and individual-level training from within a firm. This chapter overcomes this difficulty by utilizing an employer–employee dataset and considering the effects of poaching externality on training. Labour turnover is more frequent in developing countries, especially in Thailand, rather than in European countries, and so Thailand provides a unique case in understanding the issue. Here, I introduce related empirical literature on the relationship between workplace training and poaching externality. Two recent papers are important. First, Brunello and Gambarotto (2007) raise a similar question about training and poaching. They ask whether local employment density affects workplace training. One hypothesis on interpreting the results is poaching

294

Tomohiro Machikita

externality. If this effect is stronger in denser areas, there is little evidence of training. The other hypothesis is on complementarities between local knowledge spillover and skills. Local knowledge spillover based on agglomeration economies makes workplace training efficient. If these complementarities are stronger in denser areas, there is more evidence of training and the return to training is higher in denser areas. Brunello and Gambarotto (2007) test the negative effect of spatial agglomeration on workplace training, using UK regional data. The evidence from the UK suggests that training is less frequent in economically denser areas. This result suggests that the negative effect of poaching on training is stronger than raising training efficiency due to agglomeration economies. Brunello and de Paola (2008) also present a search-match training model within a local labour market framework. The magnitude of local economic density in a matching model on training has been estimated. Brunello and de Paola (2008) test whether the labour turnover in denser areas reduces workplace training, by using a dataset from Italian firms. The evidence from Italy suggests that there is less training in provinces with a higher labour market density, measured by the number of employees per square kilometre. Like Brunello and Gambarotto (2007), Brunello and de Paola (2008) shows the possibility of complementarity between local knowledge spillover, based on an agglomeration economy and training. This generates a positive correlation between the incidence of training and local density. However, this effect is smaller than poaching externality that reduces firm-provided training. Brunello and Gambarotto (2007) and Brunello and de Paola (2008) emphasize the importance of returns to training being higher in denser areas, because complementarity between local knowledge spillover and skills create more efficient workplace training. These studies succeed in detecting the effects of spatial density on training; however, the evidence on the return to training has not been fully examined. I have attempted to detect not only evidence of the poaching effect on training, but also the return to training, using individual wage data. Using individual datasets is critical to understand returns to training in each cluster. Aggregate datasets do not provide us with the information of returns to training across clusters. Finally, there are important theoretical papers on agglomeration and training. Moen and Rosén (2004) describe the efficiency of the labour market outcome in a competitive search equilibrium model, to adopt endogenous turnover and endogenous formation of human capital. The role of efficient coordination devices on training is examined. The most surprising contribution by Moen and Rosén (2004) is that they show a situation wherein the individual and total amount of training are constrained optimally when turnover is too great and there is too little investment in general training. Internal efficiency can be achieved if workers and firms agree to a long-term written contract. This result is in sharp contrast to Acemoglu’s

Clusters and Training to Expand Innovation Capability

295

(1997) finding that there is under-investment in general training due to an inability to conclude contracts between workers and prospective employers. Acemoglu (1997) emphasizes the negative effect of search friction on the training incidence. This result is also problematic for policymakers, because training subsidies are inefficient and reduce welfare. Although I have not tested the implication of Moen and Rosén’s (2004) work, I have discussed the policy implication of training subsides, based on empirical estimates, in Section 9.6. Meanwhile, it is important to detect the extent of the poaching externality. Almazan et al. (2007) have explicitly developed a model of firm-provided training with an endogenous location choice of firms. Both the formation of industrial clusters and the creation of human capital within cluster are analysed theoretically. They show the conditions of concentration and dispersion of firms, based on training costs and the sharing of training costs between workers and firms. Finally, I show the importance of studying with labour theory-based agglomeration in terms of shaping industrial cluster policy. The Flowchart Approach is one example of cutting-edge methodology to make a list of local public policy to the formation of industrial cluster. Chapters in Kuchiki and Tsuji (2005, 2008) have one’s priorities right. Unfortunately, the Flowchart Approach has not applied to the labour market. We have not fully understood the policy effects of clustering of anchor firms, supporting firms and workers on local economic development. If skills should flow locally, innovation activities may be locally emerged. Transaction of skills is now an important centre of local innovation. The labour-based agglomeration framework of this chapter helps to make a list of local public policy in terms of the labour market. Fujita (2008) shows an impressive discussion of agglomeration of innovation activity to describe circular causality between demand and supply for diverse people or ‘brain-workers’. Labour market analysis is required to understand the transaction between firms and workers. The situation of agglomeration in developing economies, especially Thailand, has also turned from formation of industrial cluster to upgrade. Brimble and Doner (2007) and Part I of this book stresses that university– industry linkages are important sources of upgrading. Kuchiki and Tsuji (2005, 2008) also emphasize the effectiveness of knowledge spillovers from university and research institute to business. University–industry linkages are also discussed in the conclusion.

9.3

Clustering firms and workplace training

This chapter focuses on the effects of agglomeration on activities of innovation capacity building. This question is central for understanding endogenous R&D and innovation mechanism within agglomeration of firms and workers. Understanding the mechanism of endogenous R&D and innovation

296 Tomohiro Machikita

plays a key role to the Flowchart Approach makes a list of local innovation policy in a logical order. In this section, I describe a simple model from which to derive empirical hypothesis. Although a large number of empirical studies have been made on firm-provided training, little is known about the effects of thick market externality on training.2 The framework of timing, location and production are fairly simplified. Take two firms locate in different thicknesses of the labour market pool, for example. Labour market thickness for workers means a higher level of turnover across firms. Workers can move across firms to seek better job matching within the thick market. On the other hand, firms can also find and hire new better workers from labour market pooling. Thick market provides positive externality for search and matching between firms and workers. If this is true, average quality of matching should be higher in thick market. Furthermore, let us assume that one firm locates at a thicker labour market while another locates at a less thick labour market. Each firm produces similar goods to compete using a worker as an input. Productivity differs quantitatively among workers. If a firm can hire a highly productive worker, the firm’s productivity becomes higher. The productive worker receives a higher wage. Timing is also simple: two periods. Each firm has two options in terms of training in the beginning of period 1: (1) the firm provides a general training programme for new employees to become productive workers; (2) the firm does not provide any training. To keep things simple, I do not distinguish between on-the-job and off-the-job training here. Productivity in the end of period 1 depends on the worker’s ex ante skill level. In the beginning of period 2, the training outcome is revealed. Another potential firm can find and hire such trained workers with substantial search-matching costs in the beginning of period 2. Training firms are under external threat of poaching their workers. Firms in thick labour markets can choose one productive worker from among many because ex ante average productivity is assumed to be higher in thick labour markets than in less thick markets. If search-matching costs are not high enough, no investment in training is the optimal strategy for firm located in thick market in terms of training.3 This basic framework suggests following the probit model of the effect of poaching on training evidence. Pr(T
1)
1P, where T is variable of training incidence when 1 equals to firm-provided training for workers while 0 equals to otherwise, P captures poaching externality within the thick labour market. This is the first testable hypothesis based on the above basic framework. The difference of the return to training between a firm’s technology or industries may also be important for the firm when it decides training incidence and intensity. The impact of

Clusters and Training to Expand Innovation Capability

297

heterogeneity in the return to training should be considered following the empirical model. Pr(T
1)
1P  , where  shows the return to training and this is industry-specific. We usually have no observational data capturing the return to training. This chapter tries to estimate it and interpret the effects of the difference of the return to training on training incidence. I assume that productivity of the individual worker is shown by ω and this is equal to the log of wage. The return to training for firm and worker is estimated by the following equation. Training incidence positively affects the firm’s productivity and the worker’s wage. ω
logW
β0  β1T  β2 X   I  u, where X and u are individual observed and unobserved characteristics. To simplify our argument, we focus on the effects of labour market poaching and the difference of the return to training across the industry in terms of empirical study.4 Testable hypotheses are summarized as follows. Primary Hypothesis: The incidence of training is not common in an industry with high turnover rates and the return to training is lower in industries with high turnover rates.

9.4

Data

9.4.1 Data source I utilize the dataset from the ADBI-KIER Employee Survey on Education and Training for Selected Manufacturing Firms in Thailand. This was also used by Ariga and Brunello (2006). They conducted a matched employer– employee survey in 20 manufacturing firms in the greater Bangkok area in Thailand in July 2001. The sample industries are four manufacturers: food processing, auto parts, hard disk drives and computer components. The main variables in the survey are firm-sponsored training, wages and individual characteristics which are difficult to observe by an employer at the hiring level. Wage and training data for 1998, 1999, 2000 and 2001 were also collected retrospectively, in July 2001. If the adoption level of new technology is different between firms in the food processing industry and firms producing hard disk drives or computer components, the provision of firm-sponsored training is also different among firms. Due to this, they gathered information not only on the industry, promotion, the employees’ career history and their family background, but also on the adoption of new technology and the training incidence.

298

Tomohiro Machikita

In July 2001, 20 firms agreed to participate in the survey: (1) five firms in food processing; (2) five firms in the auto parts industry; (3) six firms in personal computers; and (4) four firms in the hard-disk drive industry. These firms have more than 100 employees each while the hard-disk drive firms have more than 1,000 employees. The former study and this one take into account that this is not a statistically representative sample in Thailand.5 However, these are main manufacturing and exporting industries currently operating in Thailand. Because export-oriented industries face international competition, they adopt new technology and innovative human resource management policies to remain competitive. Therefore, it is reasonable to say that they provide more training more often to adapt to new technology and institutions. I explain the basic feature of our dataset. For educational attainment by their sampling criteria, the average duration of education is 12.90 years for male workers and 10.66 years for female workers. The survey sample in the manufacturing industry has a longer duration of education than the general Thai population. This suggests that educational attainment is one criterion for employment in these industries. This also suggests that younger workers dominate in this industry. The average wage is 14,386 baht per month for male workers and 9,347 baht per month for female workers. The average wage is 8,915 baht per month for production workers and 16,023 baht per month for non-production workers. 9.4.2

Summary statistics

I present certain summary statistics of important variables before moving on to empirical evidence using probit and OLS estimates in the next section. The preliminary results of this chapter are also shown here. Table 9.1 shows the summary statistics of on-the-job training (OJT) incidence in 2001, by the duration of outside labour market experience and years of tenure with the current employer. The duration of outside labour market experience is not potential experience. This does not include the duration of tenure with the current employer. Table 9.1 is divided into three occupation categories: the total, production workers, and non-production workers. Table 9.1 suggests that employees who have long experience in the outside labour market, for instance over 20 years (39.5 per cent), do not receive OJT. This is true in the case of tenure. Employees who have a long tenure with a company, for instance over 20 years (33.3 per cent), do not receive more OJT than employees who have a short tenure with the company. The provision of firm-sponsored OJT is more popular for inexperienced employees outside the current employer (60.5 per cent for 0 to 4 years and 66.5 per cent for 5 to 9 years) and short-tenure employees (64.5 per cent for 0 to 4 years and 58.3 per cent for 5 to 9 years), for example, new workers. This tendency is true for both production and non-production workers. The main difference between production and non-production workers, in receiving OJT, is the

Clusters and Training to Expand Innovation Capability

299

Table 9.1 Summary statistics of training incidence by outside labour market experience and tenure for production and non-production workers: On-the-job training incidence in 2001 (N = 1,867) Whole Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

974 525 206 86 76

0.605 0.665 0.602 0.57 0.395

0.489 0.473 0.491 0.498 0.492

844 657 304 41 21

0.645 0.583 0.612 0.512 0.333

0.479 0.493 0.488 0.506 0.483

Production Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

577 400 164 68 69

0.652 0.69 0.604 0.676 0.42

0.477 0.463 0.491 0.471 0.497

595 445 202 24 12

0.667 0.634 0.639 0.583 0.333

0.472 0.482 0.482 0.504 0.492

Non-production Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

397 124 42 18 7

0.537 0.581 0.595 0.167 0.143

0.499 0.495 0.497 0.383 0.378

249 211 102 17 9

0.59 0.474 0.559 0.412 0.333

0.493 0.501 0.499 0.507 0.5

Source: ADBI-KIER Employee Survey.

effect of the duration of outside labour market experience and the duration of tenure. Non-production workers do not receive OJT, because they have more experience in an outside company or in their current company, for example, an employee with 15 to 19 years, or over 20 years, experience and tenure. In summary, OJT is substitutable for the duration of job experience. This substitution effect is stronger for non-production workers than for production workers.6

300 Tomohiro Machikita

Table 9.2 Summary statistics of training incidence by outside labour market experience and tenure for production and non-production workers: Off-the-job training incidence in 2001 (N = 1,867) Whole Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

974 525 206 86 76

0.677 0.602 0.631 0.558 0.421

0.468 0.49 0.484 0.5 0.497

844 657 304 41 21

0.609 0.633 0.701 0.659 0.714

0.488 0.482 0.459 0.48 0.463

Production Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

577 400 164 68 69

0.652 0.558 0.604 0.529 0.391

0.477 0.497 0.491 0.503 0.492

595 445 202 24 12

0.578 0.587 0.639 0.708 0.833

0.494 0.493 0.482 0.464 0.389

Non-production Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

397 124 42 18 7

0.713 0.742 0.738 0.667 0.714

0.453 0.439 0.445 0.485 0.488

249 211 102 17 9

0.683 0.73 0.824 0.588 0.556

0.466 0.445 0.383 0.507 0.527

Source: ADBI-KIER Employee Survey.

Table 9.2 shows off-the-job training (OFFJT) data. Employees who have much experience outside their current company do not receive more OFFJT than employees who have little experience outside their current company. On the other hand, the effect of tenure on OFFJT is quite different; that is, employees who have a long tenure receive more OFFJT than those who have a short tenure. For production workers, the provision of OFFJT gradually declines as outside labour market experience grows. This is not true in the case of tenure. The provision of OFFJT shows a slight increase as

Clusters and Training to Expand Innovation Capability

301

tenure increases. The effect of outside experience and the effect of inside experience with a current company is quite an interesting difference. For non-production workers, a rapid drop in OFFJT because of experience is not stronger than it is for production workers. Employees who have much experience outside the company receive the same training as employees with little experience. On the other hand, there is a rapid drop in the provision of OFFJT for non-production workers as employee tenure increases. Of the employees who have 0 to 14 years of tenure, 82.4 per cent receive OFFJT, while under 60 per cent of employees with 15 or more years of tenure receive OFFJT. In summary, in the case of OFFJT, the effect of outside labour market experience and the effect of tenure are quite different. There is a negative relationship between the provision of OFFJT and the duration of outside experience. This relationship appears strongly for production workers. There is positive relationship between OFFJT and tenure. On the other hand, Table 9.3 presents the average level of the log of wages in July 2001 and its standard deviation for production and non-production workers, respectively. Interestingly, for production workers, there is quite a different wage profile for duration of experience in the outside labour market and tenure. Employees who have been in the outside labour market for a long time receive lower wages than employees with little experience in the outside labour market. However, employees who have a longer tenure receive higher wages than new employees. This suggests that there is a return to firm-specific human capital, or wage seniority in the firm. For non-production workers, the effect of experience in the outside labour market on the wage level is positive. More experience leads to better wages for non-production workers. The effect of tenure also impacts positively on the wage level. Finally, the turnover evidence for a sample of workers is shown in Table 9.4. Table 9.4 shows duration of education, duration of potential experience, duration of outside experience and tenure in the current firm, by industry and occupation respectively. There are four industries and two occupations: production workers and non-production workers of less than 40 years of age. The most striking evidence here is the difference in tenure for current firms between industries. Average years of tenure are one measure of turnover rates in an industry. A shorter duration means that there is a high turnover rate in the industry. Each industry forms an industrial cluster within the greater Bangkok area. The extent of the four industrial clusters is geographically limited. Higher turnover rates suggest that there are large levels of poaching externality in each industrial cluster. Average tenures are greatest in the auto parts industry, that is, 6.393 for the entire sample, 6.147 for production workers and 6.845 for non-production workers. On the contrary, the PC and HDD industries show shorter tenures in current firms. Average tenures in the PC industry are 4.816 for the entire sample, 4.639 for production workers and 5.228 for non-production workers. Average tenures

302

Tomohiro Machikita

Table 9.3 Summary statistics of log of wage in July 2001 by outside labour market experience and tenure for production and non-production workers: Log of monthly wage (N = 1,864) Whole Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

972 525 205 86 76

9.031 8.896 8.843 8.834 8.487

0.569 0.593 0.591 0.616 0.687

843 656 303 41 21

8.757 9.018 9.21 9.202 9.561

0.561 0.554 0.608 0.652 0.504

Production Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

577 400 163 68 69

8.86 8.749 8.692 8.685 8.357

0.501 0.463 0.439 0.51 0.466

594 445 202 24 12

8.596 8.827 9.059 9.053 9.508

0.464 0.41 0.536 0.438 0.465

Non-production Outside

Tenure

Years

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

0 to 4 5 to 9 10 to 14 15 to 19 Over 20

395 124 42 18 7

9.282 9.374 9.426 9.395 9.77

0.57 0.708 0.733 0.67 1.154

249 210 101 17 9

9.14 9.424 9.511 9.412 9.631

0.588 0.602 0.633 0.84 0.573

Source: ADBI-KIER Employee Survey.

in the HDD industry are 4.875 for the entire sample, 5.021 for production workers, and 4.597 for non-production workers. Evidence from the PC and HDD industries clearly suggests a shorter average tenure. There are higher turnover rates within the PC and HDD industries than within the food and auto parts industries. It is also important to note that PC and HDD industries create many jobs to deal with large demand during the survey years. We must not forget that this industry growth lowers the average tenures on average when we interpret our empirical results.

Table 9.4

Job tenure and previous experience by industry and occupation Food

Auto parts

PC

HDD

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

444 444

10.401 12.367

3.808 7.100

440 440

13.243 9.491

2.208 4.616

452 452

12.212 10.615

2.729 5.207

447 447

12.378 10.336

2.802 5.081

444

6.721

5.981

440

3.202

2.736

452

5.854

4.317

447

5.517

4.249

444

5.736

5.063

440

6.393

3.796

452

4.816

3.784

447

4.875

3.092

323 323

9.418 13.034

3.640 7.169

285 285

12.481 9.660

2.101 4.749

316 316

11.427 11.022

2.595 5.124

292 292

11.281 11.137

2.486 5.171

323

7.526

6.263

285

3.618

2.937

316

6.443

4.407

292

6.147

4.434

323

5.573

4.948

285

6.147

3.821

316

4.639

3.699

292

5.021

2.910

Whole sample Years of education Years of potential experience Years of outside experience Years of tenure Production workers Years of education Years of potential experience Years of outside experience Years of tenure

Continued

Table 9.4

Continued Food

Auto parts

PC

HDD

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

Obs

Mean

Std. dev.

121 121

13.025 10.587

2.902 6.618

155 155

14.645 9.181

1.646 4.360

136 136

14.037 9.669

2.088 5.296

154 154

14.481 8.805

2.074 4.561

121

4.570

4.518

155

2.439

2.129

136

4.485

3.773

154

4.312

3.607

121

6.174

5.355

155

6.845

3.721

136

5.228

3.957

154

4.597

3.414

Non-production workers Years of education Years of potential experience Years of outside experience Years of tenure

Note: This table reports summary statistics of years of potential experience, years of outside experience and years of tenure by industry and occupation in 2001. Source: ADBI-KIER Employee Survey.

Clusters and Training to Expand Innovation Capability

9.5

305

Results

9.5.1 Training incidence is lower in industries with higher turnover rates There are several approaches to testing the effect of agglomeration economies on workplace training. One is to compare the training incidence in denser and less dense areas, as both Brunello and Gambarotto (2007) and Brunello and de Paola (2008) apply to UK dataset and Italy dataset respectively. The other is to compare the training incidence in industrial clusters with thick and active labour markets and industrial clusters without thick and active labour markets. This chapter uses the latter approach to detect the effects of agglomeration economies on firm-provided training. The advantage of this approach is that the researchers can take into account the geographic feature of developing countries. Economic geography of developing countries, especially location of firms in Thailand, contrasts clearly with cases of the UK or Italy. Agglomeration of firms and industry in the greater Bangkok area is in strong contrast with agglomeration of firms in developed countries. Denser areas in Thailand have almost all of its industries. It is difficult for the researcher to compare the training incidence in denser and less dense areas. Alternatively, I make two different assumptions. First, I assume that the extent of the local labour market is restricted to within the greater Bangkok area. Secondly, I assume that human capital is industry-specific. These assumptions enable us to compare the differences in thickness between industry-specific local labour markets. I examine the effects of the characteristics of industry clusters on training evidence. The basic regression we estimate is that the on-the-job and off-the job training incidence in 2001 regresses to a dummy variable of industrial clusters: the food, PC and HDD industries compared to the auto parts industry, under controlling individual characteristics for workers under the age of 40 years. Informal case studies and interviews with managers suggest that training and turnover evidence within industrial clusters is not frequent for workers over the age of 40 years. This chapter utilizes case studies and restricts the estimated sample. I use the duration of the education of individual workers’ parents as a proxy of the absorption capacity for individual workers. These proxies measure the efficiency of workplace training. In Table 9.5, I systematically investigate the relevance of this proposition by estimating the industry differential on the workplace training incidence. The estimated coefficient of interest is on food (PC and HDD), an industry dummy variable that equals one if an employee belongs to a food (PC and HDD) industrial cluster in the greater Bangkok area. Column 1 of Table 9.5 estimates the basic effect of industrial clusters on the OJT incidence. The basic effect does not show the difference between industrial clusters, because omitted variables in this specification correlate to industrial cluster

Table 9.5 Industry effects on training incidence, dependent variable: Binomial OJT incidence in 2001 (Probit) (1)

(2)

(3)

(4)

(5)

(6)

(7)

Dependent: OJT incidence = 1, otherwise 0 Food PC HDD

−0.05 (0.086) −0.163 (0.085) 0.194 (0.087)*

Female

−0.079 (0.086) −0.06 (0.089) 0.159 (0.088) 0.301 (0.067)**

Years of education

−0.201 (0.091)* −0.15 (0.091) 0.136 (0.088) 0.184 (0.073)* −0.047 (0.011)**

−0.231 (0.092)* −0.193 (0.093)* 0.089 (0.089) 0.201 (0.073)** −0.051 (0.012)** −0.029 (0.008)**

−0.215 (0.092)* −0.194 (0.093)* 0.099 (0.090) 0.185 (0.074)* −0.043 (0.013)** −0.027 (0.008)** 0.127 (0.074)

−0.223 (0.093)* −0.198 (0.093)* 0.096 (0.090) 0.183 (0.074)* −0.045 (0.013)** −0.027 (0.008)** 0.128 (0.074) 0.009 (0.009)

0.837 (0.185)** 1,783

1.062 (0.197)** 1,783

0.868 (0.227)** 1,782

0.835 (0.231)** 1,782

Years of tenure Production workers Years of father’s education Years of mother’s education Constant Observations

0.319 (0.061)** 1,783

0.135 (0.073) 1,783

−0.22 (0.093)* −0.199 (0.093)* 0.093 (0.090) 0.182 (0.074)* −0.045 (0.013)** −0.027 (0.008)** 0.129 (0.075) 0.014 (0.010) −0.013 (0.009) 0.899 (0.234)** 1,782

Notes: This table lists coefficient estimates from probit regressions relating incidence to training in on-the-job training (OJT) for the whole sample in 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

Clusters and Training to Expand Innovation Capability

307

dummy variables. Column 7 of Table 9.5 estimates the basic impact of industrial clusters on OJT, while controlling the proxies of individual abilities. The results suggest that the food and PC clusters do not provide more OJT than the auto parts cluster. In columns 3 to 7 of Table 9.5, the food cluster provides OJT for workers with less than 20 percentage points, compared to the auto parts cluster. The PC cluster also provides OJT for workers with less than 19 percentage points, compared to the auto parts cluster. These provide robust estimates of the industrial cluster differential. The HDD industry provides OJT for more than approximately 9 percentage points, compared to the auto parts cluster, but this is insignificant. Other covariates indicate that duration of education and tenure have a negative effect on the OJT incidence. The difference between production workers or non-production workers does not affect the OJT incidence. Table 9.6 and Table 9.7 show the effect of industrial clusters on the OJT incidence by occupation. In Table 9.6, the OJT incidence for production workers is examined. In columns 1 to 6 of Table 9.6, the HDD cluster provides 33 to 24 percentage points more training than the auto parts cluster. There is no evidence that the food and PC clusters provide more OJT for production workers, compared to the auto parts cluster. In Table 9.7, I have investigated the effect of the differences between industrial clusters on the OJT incidence for skilled professionals. The OJT incidence for non-production workers provides a clear contrast against that of production workers. In column 1 of Table 9.7, I show the basic effect that the food cluster provides 47.7 percentage points less OJT for non-production workers, compared to the auto parts cluster. The PC cluster provides 35.8 percentage points less OJT for non-production workers, compared to the auto parts cluster. In columns 2 to 6 of Table 9.7, I present robust evidence that the food and PC industries provide less OJT for non-production workers. On the contrary, Table 9.8 presents the results of the OFFJT incidence. Column 1 of Table 9.6 suggests the basic fact that the HDD cluster does not provide OFFJT, compared to the auto parts cluster. This effect disappears when I control individual characteristics in columns 2 to 7 of Table 9.8. Tables 9.9 and 9.10 also show the OFFJT incidences for production workers and non-production workers, respectively. Column 1 of Table 9.9 presents the basic result that the HDD cluster also does not provide OFFJT for production workers, compared to the auto parts cluster. The estimates in columns 1 and 2 of Table 9.9 suggest that the HDD cluster provides less than about 24 percentage points of OFFJT for production workers, compared to the auto parts cluster. However, this effect disappears when I control the more individual characteristics in columns 3 to 6. Contrary to the results in Table 9.9 for production workers, there are not significant differences in the OFFJT incidence for non-production workers between industrial clusters.

308 Tomohiro Machikita

Table 9.6 Industry effects on training incidence for production workers, dependent variable: Binomial OJT incidence in 2001 (Probit) (1)

(2)

(3)

(4)

(5)

(6)

Dependent: OJT incidence = 1, otherwise 0 for production workers Food PC HDD

0.119 (0.105) −0.077 (0.104) 0.336 (0.110)**

Female Years of education Years of tenure

0.143 (0.105) 0.097 (0.116) 0.314 (0.110)** 0.332 (0.091)**

0.049 (0.114) 0.028 (0.121) 0.285 (0.111)** 0.259 (0.098)** −0.029

0.029 (0.115) −0.002 (0.122) 0.254 (0.112)* 0.281 (0.099)** −0.032

0.025 (0.115) −0.004 (0.122) 0.254 (0.112)* 0.28 (0.099)** −0.034

0.033 (0.115) −0.012 (0.123) 0.244 (0.112)* 0.268 (0.100)** −0.035

(0.015)*

(0.015)* −0.025 (0.010)**

(0.015)* −0.025 (0.010)** 0.005

(0.015)* −0.026 (0.010)** 0.015

(0.011)

(0.012) −0.026

Years of father’s education Years of mother’s education Constant Observations

(0.011)* 0.317 0.052 0.478 0.655 0.635 0.779 (0.076)** (0.105) (0.237)* (0.249)** (0.254)* (0.261)** 1,216 1,216 1,216 1,216 1,216 1,216

Notes: This table lists coefficient estimates from probit regressions relating incidence to training in on-the-job training (OJT) for the whole sample in 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

In summary, empirical results suggest that: (1) the OJT incidence of nonproduction workers is infrequent in industrial clusters with higher turnover rates, especially the PC cluster; and (2) the OFFJT incidence of an entire sample of production workers is not frequent in industrial clusters with higher turnover rates, namely, the HDD cluster. 9.5.2 Returns to training are lower in industries with higher turnover rates Now, I detect the reason why training is not provided in industrial clusters with higher turnover rates, for example the PC and HDD clusters compared to the auto parts cluster. One exception is the food cluster. The food cluster targets mainly the domestic market and operates labour-intensively. The opportunity cost of training instead of production is higher than in other industries.

Clusters and Training to Expand Innovation Capability

309

Table 9.7 Industry effects on training incidence for non-production workers, dependent variable: Binomial OJT incidence in 2001 (Probit) (1)

(2)

(3)

(4)

(5)

(6)

Dependent: OJT incidence = 1, otherwise 0 for Non-production workers Food PC HDD Female

−0.477 −0.553 (0.154)** (0.161)** −0.358 −0.355 (0.149)* (0.149)* −0.057 −0.098 (0.145) (0.147) 0.199 (0.115)

Years of education

−0.592 (0.164)** −0.375 (0.150)* −0.095 (0.148) 0.16 (0.119) −0.032

−0.651 (0.167)** −0.44 (0.154)** −0.176 (0.153) 0.179 (0.120) −0.048

−0.669 (0.168)** −0.449 (0.154)** −0.188 (0.153) 0.173 (0.120) −0.053

−0.661 (0.168)** −0.451 (0.154)** −0.186 (0.153) 0.162 (0.120) −0.053

(0.025)

(0.027) −0.034 (0.014)*

(0.027)* −0.034 (0.014)* 0.017

(0.027)* −0.034 (0.014)* 0.01

(0.015)

(0.016) 0.017

Years of tenure Years of father’s education Years of mother’s education Constant Observations

0.321 0.268 (0.103)** (0.107)* 566 566

0.748 (0.392) 566

(0.016) 1.202 1.164 1.095 (0.445)** (0.447)** (0.453)* 566 566 566

Notes: This table lists coefficient estimates from probit regressions relating incidence to training in on-the-job training (OJT) for production workers in 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

The PC and HDD clusters target mainly foreign markets to sell their products. There are different reasons for a lower training incidence in the food cluster and the PC/HDD clusters, compared to the auto parts cluster. Next, I present evidence that the returns to training are lower in industries with higher turnover rates. To do this, the basic regression is that the log of wages in 2001 regresses according to training duration and other covariates. The estimated coefficient of interest is the interaction terms between the training incidence in 2000 and industrial clusters: OJT length × Food; OJT length × PC; and OJT length × HDD. Table 9.11 contains striking evidence. In column 1 of Table 9.11, the variable OJT length × PC shows that returns to training are 3 percentage points lower in the PC cluster than in the auto parts cluster. Columns 3 to 7 also show robust evidence that

Table 9.8 Industry effects on training incidence, dependent variable: Binomial OFFJT incidence in 2001 (Probit) (1)

(2)

(3)

(4)

(5)

(6)

(7)

Dependent: OFFJT incidence = 1, otherwise 0 Food

−0.016 (0.087) 0.049 (0.087) −0.186 (0.086)*

−0.002 (0.087) −0.008 (0.090) −0.167 (0.086) −0.162 (0.067)*

0.404 (0.062)** 1,783

0.506 (0.075)** 1,783

PC HDD Female Years of education

0.11 (0.092) 0.072 (0.093) −0.145 (0.087) −0.056 (0.073) 0.042 (0.011)**

Years of tenure

0.135 (0.093) 0.107 (0.094) −0.108 (0.088) −0.068 (0.073) 0.045 (0.011)** 0.022 (0.008)**

Production workers

0.105 (0.094) 0.11 (0.094) −0.129 (0.088) −0.04 (0.075) 0.03 (0.012)* 0.02 (0.008)* −0.221 (0.076)**

Years of father’s education

0.092 (0.094) 0.104 (0.095) −0.134 (0.088) −0.043 (0.075) 0.027 (0.012)* 0.02 (0.008)* −0.22 (0.076)** 0.015 (0.009)

Years of mother’s education Constant Observations

−0.111 (0.183) 1,783

−0.283 (0.193) 1,783

0.048 (0.224) 1,782

−0.011 (0.228) 1,782

0.091 (0.094) 0.105 (0.095) −0.132 (0.088) −0.042 (0.075) 0.027 (0.012)* 0.02 (0.008)* −0.221 (0.076)** 0.012 (0.010) 0.008 (0.009) −0.049 (0.232) 1,782

Notes: This table lists coefficient estimates from probit regressions relating incidence to training in off-the-job training (OFFJT) for non-production workers in 2001. Columns 1 and 2 are for all occupations. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

Clusters and Training to Expand Innovation Capability

311

Table 9.9 Industry effects on training incidence for production workers, dependent variable: Binomial OFFJT incidence in 2001 (Probit) (1)

(2)

(3)

(4)

(5)

(6)

Dependent: OFFJT incidence = 1, otherwise 0 for production workers Food PC HDD

0.045 (0.104) 0.066 (0.105) −0.248 (0.105)*

Female Years of education Years of tenure Years of father’s education Years of mother’s education Constant Observations

0.039 (0.104) 0.022 (0.115) −0.242 (0.105)* −0.082 (0.090)

0.149 (0.113) 0.101 (0.120) −0.209 (0.107) 0.001 (0.098) 0.033

0.169 (0.114) 0.126 (0.121) −0.183 (0.108) −0.015 (0.098) 0.035

0.165 (0.114) 0.125 (0.121) −0.183 (0.108) −0.016 (0.098) 0.034

0.165 (0.114) 0.125 (0.121) −0.183 (0.108) −0.016 (0.098) 0.035

(0.014)*

(0.014)* 0.02 (0.009)*

(0.014)* 0.02 (0.010)* 0.005

(0.014)* 0.02 (0.010)* 0.004

(0.011)

(0.011) 0.001

(0.010) 0.299 0.366 −0.113 −0.254 −0.276 −0.282 (0.075)** (0.105)** (0.231) (0.241) (0.246) (0.252) 1,216 1,216 1,216 1,216 1,216 1,216

Notes: This table lists coefficient estimates from probit regressions relating incidence to training in off-the-job training (OFFJT) for production workers in 2001. Robust standard errors in parentheses.* significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

returns to training are 1.8 percentage points to 2.2 percentage points lower in the PC cluster than in the auto parts cluster, after controlling individual characteristics. The returns to training in the HDD cluster are also negative and consistent with the case of a PC cluster with a high turnover rate. However, this is not significant in columns 1 to 7 of Table 9.11. Table 9.12 and Table 9.13 present the results of the returns to OJT duration in different industries, for production and non-production workers, respectively. In columns of 1 to 3 of Table 9.12, the coefficient of OJT length×Food for production workers suggests that this is 4.5 to 2.9 percentage points lower, compared to the auto parts cluster. The coefficients of OJT length × PC and OJT length × HDD for production workers are negative and insignificant. In column 2 of Table 9.13, the coefficient of OJT length × Food for

312

Tomohiro Machikita

Table 9.10 Industry effects on training incidence for non-production workers, dependent variable: Binomial OFFJT incidence in 2001 (Probit) (1)

(2)

(3)

(4)

(5)

(6)

Dependent: OFFJT incidence = 1, otherwise 0 for non-production workers Food PC HDD Female

−0.102 (0.161) 0.065 (0.159) −0.063 (0.152)

−0.087 (0.167) 0.064 (0.159) −0.054 (0.154) −0.04 (0.120)

Years of education

−0.075 (0.170) 0.07 (0.160) −0.056 (0.154) −0.027 (0.124) 0.01

−0.047 (0.173) 0.11 (0.162) −0.009 (0.158) −0.035 (0.124) 0.02

−0.082 (0.174) 0.091 (0.163) −0.034 (0.159) −0.048 (0.125) 0.009

−0.07 (0.174) 0.091 (0.163) −0.03 (0.159) −0.06 (0.125) 0.009

(0.026)

(0.027) 0.02 (0.015)

(0.027) 0.02 (0.015) 0.035

(0.028) 0.02 (0.015) 0.025

(0.016)*

(0.018) 0.023

Years of tenure Years of father’s education Years of mother’s education Constant Observations

0.61 0.621 0.467 (0.108)** (0.113)** (0.408) 566 566 566

0.19 (0.452) 566

(0.018) 0.108 0.008 (0.458) (0.464) 566 566

Notes: This table lists coefficient estimates from probit regressions relating incidence to training in off-the-job training (OFFJT) for non-production workers in 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

non-production workers is 4 percentage points lower, compared to the auto parts cluster. Finally, I show the basic impact of the returns to OFFJT duration for higher turnover rates and lower turnover rates in Table 9.14. The estimated coefficient of interest is the OFFJT length × Food, OFFJT length × PC, and OFFJT length × HDD. In columns 1 to 7 of Table 9.14, the coefficient of OFFJT length × Food is positive and significant. On the contrary, the estimated coefficient of OFFJT length × PC and OFFJT length × HDD is insignificant. These robust estimates suggest that the returns to OFFJT duration are higher in the food cluster while these disappear in the PC and HDD clusters, compared to the auto parts cluster. Table 9.15 and Table 9.16 show evidence that the returns to training differ in industrial clusters, for

Table 9.11 Effects of interactions between OJT length and industry on log of wage, dependent variable: Log of wage in July 2001 (OLS) (1)

(2)

(3)

(4)

(5)

(6)

(7)

Log of wage per month in 2001 OJT length in 2000 Food PC HDD OJT length in 2000  Food OJT length in 2000  PC OJT length in 2000  HDD Female Years of education Years of tenure Production workers

0.045 (0.008)** −0.337 (0.052)** 0.017 (0.056) −0.053 (0.054) 0.026 (0.014) −0.03 (0.012)* −0.018 (0.012)

0.031 (0.007)** −0.33 (0.048)** −0.161 (0.057)** −0.021 (0.048) 0.032 (0.012)** −0.02 (0.011) −0.012 (0.010) −0.423 (0.037)**

0.029 (0.007)** −0.084 (0.048) −0.008 (0.053) 0.035 (0.045) 0.009 (0.011) −0.02 (0.010)* −0.016 (0.009) −0.232 (0.036)** 0.074 (0.006)**

0.027 (0.005)** −0.013 (0.044) 0.086 (0.045) 0.125 (0.039)** 0.005 (0.010) −0.022 (0.008)** −0.015 (0.008) −0.258 (0.033)** 0.084 (0.005)** 0.054 (0.003)**

0.024 (0.005)** −0.04 (0.045) 0.08 (0.043) 0.112 (0.039)** 0.011 (0.010) −0.018 (0.008)* −0.012 (0.008) −0.235 (0.032)** 0.073 (0.006)** 0.052 (0.003)** −0.182 (0.033)**

0.024 (0.005)** −0.051 (0.045) 0.076 (0.043) 0.11 (0.039)** 0.012 (0.010) −0.018 (0.008)* −0.013 (0.008) −0.235 (0.032)** 0.071 (0.006)** 0.052 (0.003)** −0.177 (0.032)**

0.024 (0.005)** −0.051 (0.045) 0.075 (0.043) 0.11 (0.039)** 0.011 (0.010) −0.018 (0.008)* −0.013 (0.008) −0.233 (0.032)** 0.071 (0.006)** 0.052 (0.003)** −0.177 (0.032)** Continued

Table 9.11 Continued (1)

(2)

(3)

(4)

(5)

(6)

(7)

Log of wage per month in 2001 0.012 (0.004)**

Years of father’s education Years of mother’s education Constant Observations R-squared

8.884 (0.040)** 1,074 0.15

9.209 (0.050)** 1,074 0.27

8.127 (0.087)** 1,074 0.39

7.666 (0.086)** 1,074 0.53

7.937 (0.097)** 1,073 0.54

7.874 (0.099)** 1,073 0.55

0.01 (0.004)** 0.004 (0.004) 7.852 (0.103)** 1,073 0.55

Notes: This table lists coefficient estimates from OLS relating log of wage in July 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

Clusters and Training to Expand Innovation Capability

315

Table 9.12 Effects of interactions between OJT length and industry on log of wage for production workers, dependent variable: Log of wage in July 2001 (OLS) (1)

(2)

(3)

(4)

(5)

(6)

Log of wage per month in 2001 for production workers OJT length in 2000 Food PC HDD OJT length in 2000  Food OJT length in 2000  PC OJT length in 2000  HDD Female

0.025

0.018

0.017

(0.010)* −0.22 (0.050)** 0.03 (0.054) −0.012 (0.048) 0.045

(0.008)* −0.25 (0.049)** −0.14 (0.059)* −0.007 (0.045) 0.045

(0.008)* −0.066 (0.054) −0.01 (0.058) 0.047 (0.044) 0.029

(0.006)** 0.008 (0.048) 0.093 (0.048) 0.123 (0.037)** 0.015

(0.006)** 0.004 (0.048) 0.092 (0.048) 0.126 (0.037)** 0.014

(0.006)** 0.004 (0.048) 0.092 (0.048) 0.126 (0.037)** 0.014

(0.017)** −0.005

(0.015)** −0.001

(0.014)* −0.003

(0.011) −0.012

(0.011) −0.013

(0.012) −0.013

(0.013) −0.013

(0.012) −0.011

(0.011) −0.013

(0.008) −0.015

(0.008) −0.016

(0.008) −0.016

(0.013)

(0.011) −0.308 (0.045)**

(0.010) −0.181 (0.043)** 0.051

(0.009) (0.009) (0.009) −0.193 −0.194 −0.193 (0.040)** (0.040)** (0.040)** 0.061 0.06 0.06

(0.007)**

(0.006)** 0.052 (0.003)**

Years of education Years of tenure Years of father’s education

0.021

0.021

(0.006)** (0.006)** 0.052 0.052 (0.003)** (0.003)** 0.005 0.005 (0.003)

Years of mother’s education

0.021

(0.003) 0.001 (0.003)

Constant Observations R-squared

8.742 9.022 8.295 (0.036)** (0.058)** (0.104)** 767 767 767 0.1 0.18 0.27

7.84 7.813 7.81 (0.099)** (0.099)** (0.103)** 767 767 767 0.47 0.47 0.47

Notes: This table lists coefficient estimates from OLS relating log of wage in July 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1% Source: ADBI-KIER Employee Survey.

316

Tomohiro Machikita

Table 9.13 Effects of interactions between OJT length and industry on log of wage for non-production workers, dependent variable: Log of wage in July 2001 (OLS) (1)

(2)

(3)

(4)

(5)

(6)

Log of wage per month in 2001 for non-production workers OJT length in 2000 Food PC HDD OJT length in 2000  Food OJT length in 2000  PC OJT length in 2000  HDD

0.032

0.027

0.025

0.021

0.019

0.019

(0.012)** −0.651 (0.117)** −0.055 (0.114) −0.153 (0.119) 0.037

(0.011)* −0.555 (0.113)** −0.074 (0.114) −0.086 (0.114) 0.041

(0.010)** −0.32 (0.102)** −0.01 (0.098) −0.088 (0.102) 0.011

(0.008)* −0.198 (0.098)* 0.096 (0.084) 0.064 (0.096) 0.018

(0.008)* −0.226 (0.100)* 0.084 (0.083) 0.021 (0.094) 0.021

(0.008)* −0.231 (0.101)* 0.07 (0.082) 0.008 (0.093) 0.023

(0.025) −0.029

(0.020)* −0.029

(0.025) −0.025

(0.020) −0.019

(0.022) −0.018

(0.021) −0.017

(0.020) 0.005

(0.020) 0.005

(0.021) 0.001

(0.015) −0.001

(0.017) 0.004

(0.016) 0.006

(0.017)

(0.016) −0.317 (0.066)**

(0.014) −0.154 (0.061)* 0.114

(0.013) −0.232 (0.059)** 0.137

(0.012) (0.012) −0.225 −0.226 (0.057)** (0.056)** 0.132 0.133

(0.014)**

(0.014)** 0.056 (0.007)**

(0.014)** 0.054 (0.007)** 0.026

(0.014)** 0.055 (0.008)** 0.019

(0.008)**

(0.008)* 0.016

Female Years of education Years of tenure Years of father’s education Years of mother’s education Constant Observations R-squared

(0.009) 9.303 9.424 7.729 7.043 6.946 6.873 (0.082)** (0.085)** (0.205)** (0.220)** (0.217)** (0.229)** 306 306 306 306 306 306 0.21 0.27 0.4 0.51 0.53 0.54

Notes: This table lists coefficient estimates from OLS relating log of wage in July 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

Table 9.14 Effects of interactions between OFFJT length and industry on log of wage, dependent variable: Log of wage in July 2001 (OLS) (1)

(2)

(3)

(4)

(5)

(6)

(7)

Log of wage per month in 2001 OFFJT length in 2000 Food PC HDD OFFJT length in 2000  Food OFFJT length in 2000  PC OFFJT length in 2000  HDD Female Years of education Years of tenure Production workers

0.037 (0.039) −0.589 (0.088)** −0.099 (0.064) −0.148 (0.065)* 0.302 (0.088)** 0.023 (0.043) −0.021 (0.044)

0.004 (0.026) −0.522 (0.084)** −0.266 (0.056)** −0.101 (0.050)* 0.29 (0.088)** 0.033 (0.032) 0.009 (0.030) −0.503 (0.035)**

0.002 (0.024) −0.252 (0.069)** −0.084 (0.052) −0.042 (0.046) 0.224 (0.065)** 0.018 (0.030) 0.004 (0.028) −0.283 (0.033)** 0.089 (0.005)**

0.012 (0.016) −0.121 (0.061)* −0.017 (0.042) 0.072 (0.038) 0.131 (0.054)* 0.001 (0.024) −0.015 (0.020) −0.293 (0.030)** 0.1 (0.005)** 0.054 (0.003)**

0.014 (0.016) −0.151 (0.060)* −0.015 (0.040) 0.062 (0.037) 0.144 (0.053)** 0.002 (0.022) −0.018 (0.020) −0.266 (0.030)** 0.088 (0.005)** 0.052 (0.003)** −0.183

0.013 (0.016) −0.167 (0.060)** −0.02 (0.040) 0.062 (0.037) 0.149 (0.053)** 0.004 (0.022) −0.018 (0.020) −0.268 (0.029)** 0.085 (0.005)** 0.052 (0.003)** −0.179

0.013 (0.016) −0.169 (0.060)** −0.02 (0.040) 0.062 (0.037) 0.151 (0.054)** 0.004 (0.023) −0.017 (0.019) −0.269 (0.029)** 0.085 (0.005)** 0.052 (0.003)** −0.179

(0.030)**

(0.030)**

(0.030)** Continued

Table 9.14

Continued (1)

(2)

(3)

(4)

(5)

(6)

(7)

Log of wage per month in 2001 0.012 (0.004)**

Years of father’s education Years of mother’s education Constant Observations R-squared

9.086 (0.053)** 1,249 0.07

9.429 8.114 7.606 (0.048)** (0.082)** (0.080)** 1,249 1,249 1,249 0.23 0.4 0.53

7.876 (0.091)** 1,248 0.55

7.826 (0.092)** 1,248 0.55

0.009 (0.004)* 0.008 (0.004)* 7.801 (0.093)** 1,248 0.56

Notes: This table lists coefficient estimates from OLS relating log of wage in July 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

Clusters and Training to Expand Innovation Capability

319

Table 9.15 Effects of interactions between OFFJT length and industry on log of wage for production workers, dependent variable: Log of wage in July 2001 (OLS) (1)

(2)

(3)

(4)

(5)

(6)

Log of wage per month in 2001 for production workers OFFJT length in 2000 Food PC HDD OFFJT length in 2000  Food OFFJT length in 2000  PC OFFJT length in 2000  HDD

0.273

0.236

(0.085)** 0.404 (0.088)** 0.391 (0.084)** 0.286 (0.081)** −0.217

(0.089)** 0.415 (0.087)** 0.207 (0.088)* 0.306 (0.083)** −0.207

(0.087)* −0.249

0.112

0.114

0.115

(0.068)** 0.192 (0.077)* 0.145 −0.074 0.157 (0.072)* −0.172

(0.056)* 0.071 (0.068) 0.097 (0.064) 0.136 (0.063)* −0.101

(0.055)* 0.079 (0.068) 0.106 (0.063) 0.147 (0.063)* −0.102

(0.055)* 0.081 (0.067) 0.109 (0.063) 0.15 (0.062)* −0.103

(0.090)* −0.237

(0.070)* −0.194

(0.059) −0.094

(0.058) (0.058) −0.095 −0.096

(0.092)** −0.272

(0.091)** −0.236

(0.071)** −0.203

(0.059) −0.124

(0.058) (0.058) −0.125 −0.126

(0.086)**

(0.089)** −0.405 (0.047)**

(0.068)** −0.243 (0.044)** 0.065

(0.057)* (0.056)* −0.25 −0.25 (0.040)** (0.040)** 0.076 0.074

(0.056)* −0.25 (0.040)** 0.074

(0.006)**

(0.006)** 0.053 (0.003)**

(0.006)** 0.053 (0.003)** 0.006

Female Years of education Years of tenure Years of father’s education

0.19

(0.006)** 0.053 (0.003)** 0.008

(0.004)* (0.004) 0.007

Years of mother’s education

(0.004) Constant Observations R-squared

8.466 (0.074)** 843 0.06

8.806 8.094 7.725 7.679 7.654 (0.085)** (0.090)** (0.083)** (0.084)** (0.087)** 843 843 843 843 843 0.18 0.3 0.47 0.48 0.48

Notes: This table lists coefficient estimates from OLS relating log of wage in July 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

320 Tomohiro Machikita

Table 9.16 Effects of interactions between OFFJT length and industry on log of wage for non-Production workers, dependent variable: Log of wage in July 2001 (OLS) (1)

(2)

(3)

(4)

(5)

(6)

Log of wage per month in 2001 for non-production workers OFFJT length in 2000 Food PC HDD OFFJT length in 2000  Food OFFJT length in 2000  PC OFFJT length in 2000  HDD Female

0.05

0.03

0.032

0.007

0.001

(0.054) −0.866 (0.138)** −0.178 (0.103) −0.119 (0.113) 0.437

(0.055) −0.736 (0.141)** −0.204 (0.102)* −0.062 (0.106) 0.429

(0.046) (0.034) −0.498 −0.331 (0.124)** (0.117)** −0.113 −0.073 (0.092) (0.073) −0.08 0.038 (0.091) (0.080) 0.332 0.257

(0.037) (0.036) −0.371 −0.368 (0.119)** (0.118)** −0.099 −0.101 (0.075) (0.075) 0.026 0.023 (0.081) (0.081) 0.278 0.28

(0.123)** 0.001

(0.128)** 0.019

(0.105)** 0.005

(0.100)* 0.023

(0.105)** 0.036

(0.106)** 0.035

(0.062)

(0.061)

(0.053)

(0.041)

(0.044)

(0.044)

−0.031

−0.012

0.008

0.012

0.011

0.013

(0.069)

(0.064) −0.329 (0.059)**

Years of education Years of tenure Years of father’s education

(0.050) (0.038) −0.202 −0.243 (0.053)** (0.049)** 0.114 0.138

(0.041) (0.040) −0.248 −0.252 (0.047)** (0.047)** 0.132 0.132

(0.011)**

(0.010)** (0.010)** 0.052 0.053 (0.007)** (0.007)** 0.018 0.014

(0.010)** 0.052 (0.006)**

(0.007)* Years of mother’s education Constant Observations R-squared

0.001

(0.008) 0.008

(0.007) 9.466 9.58 7.864 7.19 7.159 7.128 (0.084)** (0.083)** (0.172)** (0.172)** (0.170)** (0.172)** 405 405 405 405 405 405 0.13 0.19 0.36 0.47 0.48 0.49

Notes: This table lists coefficient estimates from OLS relating log of wage in July 2001. Robust standard errors in parentheses. * significant at 5%; ** significant at 1%. Source: ADBI-KIER Employee Survey.

Clusters and Training to Expand Innovation Capability

321

production and non-production workers, respectively. The most striking and consistent evidence of the theory in Table 9.15 is that the coefficients of OFFJT length × PC and OFFJT length × HDD for production workers are negatively significant in columns 1 to 3 of Table 9.15. The coefficient of OFFJT length × Food is estimated to be negatively significant. The magnitude of the returns to training is also important. The return to training is 21.7 to 17.2 percentage points lower in the food cluster, compared to the auto parts cluster. On the other hand, the return to training is 24.9 to 19.4 percentage points lower in the PC cluster and 27.2 percentage points lower in the HDD cluster. In columns 4 to 6 of Table 9.15, robust evidence suggests that the coefficient of OFFJT length × HDD is negatively significant under a controlling measurement of absorption capacity. The empirical results for non-production workers are shown in Table 9.16. In columns 1 to 7 of Table 9.16, the estimated coefficient of OFFJT length × Food is positively significant. The coefficients of OFFJT length × PC and OFFJT length × HDD for non-production workers are insignificant. In the next section, I will use these empirical results to derive and evaluate the policy implications of the model. In summary, robust empirical results suggest that the return to OJT length is lower in the PC cluster for the entire sample and the return to OFFJT length is lower for production workers in the HDD cluster.

9.6 Summary and conclusions I summarize the three main empirical results as follows. First, the OJT incidence for non-production workers is infrequent in industrial clusters with higher turnover rates, especially the PC cluster. Secondly, the OFFJT incidence for the entire sample and for production workers is infrequent in industrial clusters with higher turnover rates, namely, the HDD cluster. Finally, robust empirical results suggest that the return to OJT duration is lower in the PC cluster for the entire sample and the return to OFFJT duration is lower for production workers in the HDD cluster. These results help us to consider why the training incidence is infrequent in industries with high turnover rates. One possible reason is that poaching externality reduces training incidence in spite of the level of return to training. The return to training is also lower in HDD and PC cluster. The results from HDD and PC clusters suggest this interpretation. The results from food and auto clusters are relatively difficult to solve. Two types of clusters have similar characteristics: (1) higher level of the estimated return to training; (2) lower level of the labour turnover. Both of these predict that the training incidence is frequent. This should be considered further. Another possible reason is that the returns to training are lower in industrial clusters with high turnover rates. We have no definite information on the causal effect of turnover

322 Tomohiro Machikita

on returns to training. I cannot say for certain the extent of the negative effects of labour turnover on returns to training. Theoretical developments are needed to explain the mechanism of reducing returns to training due to labour turnover. I would like to emphasize there is complementary changes involving labour turnover due to thick markets and return to training that is the key firm-provided training. This is an important fact within the findings of this chapter. There are some policy implications and policy suggestions of upgrading industrial clusters based on empirical results presented here. First, clustering firms by itself does not always automatically lead to industrial upgrading. If firms need to accumulate adaptation-type (trial and error-type) human capital for future survival, workplace training is an appropriate strategy. Poaching externality within an industrial cluster reduces the benefit of training incidence and increases labour mobility and high-quality matching. That is why there is little evidence of firm-provided training and the return to training is lower in industrial clusters with high turnover rates, or the pressure of poaching externality. This is the main effect of clustering on innovation capacity building, which is emphasized throughout this chapter. Secondly, this chapter does not suggest that local authorities should prevent poaching and employee mobility between firms. Lowering the cost of adopting new technology reduces training costs for incumbent workers. Active labour market policy stimulates not only poaching externality but also knowledge spillover and matching. This stimulates the incidence of firm-provided training because the returns to training become higher within firms have higher level of knowledge spillovers.7 Thirdly, empirical results suggest that it is difficult to support simple active labour market policy without evaluating the employers’ investment in on-the-job and offthe-job training. Poaching externality reduces training benefits for employers who do not earn future rent on their trained employees. The most important implication here is evaluating the different returns to training by the degree of market thickness. Finally, this chapter has a policy suggestion that complements lower levels of firm-provided training with subsidy training evidence in university or technical college.8 To implement this, we need to know the effectiveness of general training between university–industry linkages in production lines or the workplace. More empirical examinations are urgently needed to shape the Flowchart Approach to the formation and upgrading industrial cluster. There are several remaining issues. First, I have not tracked employee mobility between former and new employers within the same industrial cluster, due to a lack of data. Secondly, I have not examined the complementarities between matching externality and training. These remaining issues can raise another policy implication for the labour-based theory of agglomeration and innovation.

Clusters and Training to Expand Innovation Capability

323

Notes 1. There is a large discrepancy between, and debate on, two major local policies. Frequent job and labour turnover affects knowledge spillover between companies and the low level of human capital accumulation in the workplace. On the other hand, innovative human resource management policy (hereafter, HRMP) in a company attracts good workers for training, while the return to training and adoption of innovative HRMP does not spillover in different firms. However, location theory and its application to the industrial cluster policy, such as that of Fujita and Thisse (2002), the capacity building approach by Kuchiki and Tsuji (2005), and the most recent critiques on industrial clusters by Duranton (2007), have not suggested how we can shape active labour market policy and innovative HRMP into evidence-based policy-making in the industrial clusters. One approach is to learn from micro-econometric study. An active labour market policy based on micro-econometric evidence changes the way we think about heterogeneity, location of available jobs and the location of human capital. 2. Recent workplace training literature is summarized very clearly by Asplund (2005) and Leuven (2005). 3. Acemoglu and Pischke (1998) provide the effects of poaching externality on firmprovided training using the framework of competitive and imperfect competitive market. The competitive market assumption relates to Becker’s prediction without labour market friction. The assumption of imperfect competitive market predicts that even general training is provided by firm to exploit the return to productivity changes based on training. Empirical results of this chapter suggests that even all firms provides on-the-job and off-the job training, training intensity and returns to training still differ across labour markets. 4. This framework simply assumes that an employer cannot perfectly observe the productivity of outside workers or new employees. Richer hypothesis requires that the employer discovers true productivity by observing the output when an employee starts the production process. Thus, an employer provides training to an employee based on productivity information set. Therefore, by estimating the effect of pre-market abilities on training, the length of tenure is more important in determining the provision of firm-sponsored training than labour market experience is. For the empirical implementation of testing the hypothesis and for estimating the effect of pre-market ability on the provision of firm-sponsored training, the effect of tenure is larger than the effect of labour market experience. If information is completely private, we could expect that returns to schooling and pre-market ability would be affected only by experience in current firms and not by previous labour market experience. Specifically, if the effect of interaction terms between education and previous labour market experience, before entering current firms, is insignificant, then the public learning assumption is unsupported. I have used the tenure in current firms and previous labour market experience to test the public learning model on the provision of firm-sponsored on-the-job and off-the-job training. This richer hypothesis should be incorporated into the agglomeration model. 5. Ariga and Brunello (2006) focus on the relationship between education and the provision of firm-sponsored training, on-the-job and off-the-job training. If there is a positive relationship between the level of education and the provision of training, then we can interpret complementarities between both the formation of human capital and firm provided training. Ariga and Brunello (2006) find: (1) a negative

324

Tomohiro Machikita

and statistically significant relationship between educational attainment and onthe-job training (OJT); and (2) a positive and statistically significant relationship between educational attainment and off-the-job training (OFFJT). The main variables used by Ariga and Brunello (2006) are education, training and wages. 6. It is very difficult to distinguish on-the-job (OJT) and off-the-job training (OFFJT). For the provision of firm-sponsored training, Ariga and Brunello (2006) define OJT that is administered in the workplace by trainers such as senior employees, supervisors or instructors. Employers expect this to improve the job performance of daily tasks. Learning by doing and learning from others on the job are not included in this OJT definition. These variables are impossible to define or measure, so Ariga and Brunello (2006), and I, do not distinguish these from regular work. Tenure can be used to capture such on-the-job learning behaviour. On the other hand, OFFJT is done by an outside company or professional training centre. This is provided in a more formal manner than the incidence of OJT in the workplace. An employer expects this to improve the general type of human capital and skills, to deal with disequilibria in the workplace. Using their sample criterion, OJT was received by 55 per cent of male workers and 67 per cent of female workers in the 2001 survey period. OFFJT was received by 67 per cent of male workers and 58 per cent of female workers. Note that Ariga and Brunello (2006) constructed not only the training incidence, but also training intensity, which is defined by the average duration (hours) of training per month. In the case of OJT, training intensity was 2.62 for male workers and 3.18 for female workers. For OFFJT, training intensity was 1.57 for male workers and 0.87 for female workers. This suggests male workers receive less OJT and more OFFJT than female workers. An employer provides more OJT for production workers and more OFFJT for non-production workers. The gender difference in the provision of training may be reflected by the gender difference in occupational distribution in the workplace. 7. On the other hand, investing in the policy of on-the-job human capital formation inside the firm, for example, an innovative human resources management policy (HRMP) is determined by a firm’s internal returns to training based on industryspecific effects and poaching externality. This is also suggested by Ichniowski et al. (1997) and Ichniowski and Shaw (2003). 8. Almazan et al. (2007) also recommend this local public policy to develop university–industry linkage.

References Acemoglu, Daron (1997) ‘Training and Innovation in an Imperfect Labor Market.’ Review of Economic Studies, Vol. 64, pp. 445–464. Acemoglu, Daron and Jorn Steve Pischke (1998) ‘Why Do Firms Train? Theory and Evidence.’ Quarterly Journal of Economics, Vol. 113, pp. 79–119. Almazan, Andres, Adolfo de Motta, and Sheridan Titman (2007) ‘Firm Location and the Creation and Utilization of Human Capital.’ Review of Economic Studies, Vol. 74, No. 4, pp. 1305–1327. Ariga, Kenn and Giorgio Brunello (2006) ‘Are Education and Training Always Complements? Evidence from Thailand.’ Industrial Labor Relations Review, Vol. 59, No. 4, pp. 613–629. Asplund, Rita (2005) ‘The Provision and Effects of Company Training: A Brief Review of the Literature,’ mimeo.

Clusters and Training to Expand Innovation Capability

325

Brimble, Peter and Richard F. Doner (2007) ‘University–Industry Linkages and Economic Development: The Case of Thailand.’ World Development, Vol. 35, No. 6, pp. 1021–1036. Brunello, Giorgio and Francesca Gambarotto (2007) ‘Do Spatial Agglomeration and Local Labor Market Competition affect Employer-provided Training? Evidence from the UK.’ Regional Science and Urban Economics, Vol. 37, No. 2, pp. 1–21. Brunello, Giorgio and Maria De Paola (2008) ‘Training and Economic Density: Some Evidence form Italian Provinces.’ Labour Economics, Vol. 15, No. 1, pp. 118–140. Duranton, Giles (2007) ‘ “California Dreamin”: The Feeble Case for Cluster Policies,’ mimeo. Fujita, Masahisa (2008)’Formation and Growth of Economic Agglomerations and Industrial Clusters: A Theoretical Framework from the Viewpoint of Spatial Economics,’ in The Flowchart Approach to Industrial Cluster Policy, Akifumi Kuchiki and Masatsugu Tsuji Eds, Basingstoke: Palgrave Macmillan, pp. 18–37. Fujita, Masahisa and Jacques F. Thisse (2002) Economics of Agglomeration: Cities, Industrial Location, and Regional Growth. Cambridge: Cambridge University Press. Ichniowski, Casey and Kathryn Shaw (2003) ‘Beyond Incentive Pay: Insiders’ Estimates of the Value of Complementary Human Resource Management Practices.’ Journal of Economic Perspectives, Vol. 17, No. 1, pp. 155–180. Ichniowski, Casey, Kathryn Shaw, and Giovanna Prennushi (1997) ‘The Effects of Human Resource Management Practices on Productivity: A Study of Steel Finishing Lines.’ American Economic Review, Vol. 87, No. 3, pp. 291–313. Kuchiki, Akifumi and Masatsugu Tsuji (2005) Industrial Clusters in Asia: Analysis of Their Competition and Cooperation. Basingstoke: Palgrave Macmillan. —— (2008) The Flowchart Approach to Industrial Cluster Policy. Basingstoke: Palgrave Macmillan. Leuven, Edwin (2005) ‘The Economics of Private Sector Training: A Survey of the Literature.’ Journal of Economic Survey, Vol. 19, No. 1, pp. 91–111. Markusen, Ann(1996)’Sticky Places in Slippery Space: A Typology of Industrial Districts.’ Economic Geography, Vol. 72, No. 3, pp. 293–313. Moen, Espen R. and Asa Rosén (2004) ‘Does Poaching Distort Training?’ Review of Economic Studies, Vol. 71, No. 4, pp. 1143–1162.

10 Innovation as a Driver for Building an Oil & Gas Industrial Cluster in Rio de Janeiro, Brazil Antonio José Junqueira Botelho and Glaudson Mosqueira Bastos

10.1

Introduction

The goal of this chapter is to discuss the role of innovation in the evolution of the oil and gas exploration (O&G) sector in the state of Rio de Janeiro (RJ). This chapter analyses, in the flowchart framework, the formation of a local innovation system in the oil & gas industry in the state of Rio de Janeiro, centred on research universities and research labs of the Brazilian state O&G company PETROBRAS – the largest exploration and production (E&P) player in Brazil with 109 production platforms (77 fixed; 32 floating) and a daily production of about 1.8 million barrels of oil a day. It proposes an specification of the conditions for innovation to impact on the establishment of an innovation-led industrial cluster in the oil & gas sector in the RJ metropolitan region and/or the transformation of the broader industrial agglomeration in the axis of the cities of Rio de Janeiro – headquarters of the demand anchor firm PETROBRAS, and the city of Macaé – the oil & gas sea exploration hub, into an innovation-based industrial cluster. In order to meet these objectives, the chapter first discusses the development of O&G production in Brazil, and in the state of RJ in particular. Next, it reviews and charts the future development of an O&G industrial agglomeration into a (proto) cluster in Campos Basin region of RJ. There is a group of organizations and government agencies funding a support network named REDE PETRO BC to assist PETROBRAS to increase outsourcing of products and services from third parties among 89 small suppliers.1 PETROBRAS’ catalytic role in inducing an industrial cluster policy in Brazil is partly due to the fact that it is controlled by the Brazilian government. Thus, besides the economic rationale argument for PETROBRAS to promote the development of an innovation-led regional cluster in the state of Rio de Janeiro, PETROBRAS is also led to collaborate with the government’s innovation and cluster policies for industrial development of SMES, part 326

Oil & Gas Cluster in Rio de Janeiro 327

Table 10.1

Evolution of O&G industry suppliers

Year

Jan.

Feb.

Mar.

Oct.

Nov.

Dec.

2000 2001 2002 2003 2004 2005 2006 2007

– 240 535 688 887 1,100 1,310 1,470

– 266 535 695 916 1,142 1,310 1,470

– 292 560 713 916 1,157 1,328 1,470

112 476 662 770 1,087 1,272 1,428 1,576

171 516 681 847 1,100 1,272 1,428

210 535 681 874 1,100 1,310 1,470

Source: ONIP (2007).

Table 10.2 South-east region O&G industry suppliers South-east ES

MG

RJ

SP

Sub-total

Total

137

96

568

545

1,328

1,576

Source: ONIP (2007).

of Brazil’s 2004–2007 Development Plan (also known as PPA),2 based on the promotion of innovation and of cooperation between enterprises and research institutions. The following section presents the RJ state local system of innovation and the O&G sector main player, PETROBRAS, innovation structure and strategy; and the local and national policies for innovation, particularly for the O&G sector. Finally, the chapter discusses the potential and problems for innovation to consolidate the emerging cluster, suggesting policy orientations. The O&G sector is extremely diverse in its industrial and service needs, being composed by a large number of suppliers, domestic and foreign, large and small and a myriad of subcontractors. As shown in Table 10.1, following the industry’s spectacular growth in the past few years, the number of suppliers has tripled since 2000, reaching, in October 2007, 1,576 suppliers registered with the bridge institution National Oil Industry Organization (Organização Nacional da Indústria do Petróleo: ONIP). The south-east region, the country’s most developed area, contains the majority of the industry (85 per cent), with the state of Rio de Janeiro alone accounting for 36 per cent, followed closely by the State of São Paulo, the country’s most industrialized area (Table 10.2). Within the state of Rio

328

A. J. J. Botelho and G. M. Bastos

Table 10.3

O&G industry suppliers by service groups

Service group Naval construction, maintenance and repair Studies and projects Industrial assembly and installation Industrial maintenance Construction Oil well exploration, drilling and production services Institutional relations services Transportation services (goods and people) General services Logistic services Specialized technical services

Suppliers 163 462 325 501 228 75 55 104 317 29 736

Source: ONIP (2007).

de Janeiro, the Rio de Janeiro Metropolitan Area3 has 454 suppliers, led by Rio de Janeiro with 381, followed far behind by the Campos Basin Area4 with 89, led by Macaé with 62 and Campos dos Goytacazes with 21. Britto (2007) identified a local production system (APL) in the Macaé region with 144 firms employing 22,517 people, and a wage turnover of R$57.5 million, the state’s largest followed by APLs5 in telecommunications (R$47.4 million and 20,000 employees) and IT (R$37.5 million and 19,000 employees), both in the city of Rio de Janeiro. ONIP classifies registered service suppliers by service groups. It also details supplies in over 100 classes of materials. It also identifies strategic gods by families of materials, which comprise a set of similar strategic goods represented by a common code. A family can be strategic or not. There are currently about 50-plus families of materials and services groups (Table 10.3).

10.2 Brazil oil production According to the BP Statistical Review of World Energy 2007,6 there’s a stability of oil proven reserves7 of approximately 1.20 trillion barrels at the end of 2006. About 75 per cent of total world reserves are found in 11 countries, members of OPEC. Reserves of the Middle East alone correspond to approximately 62 per cent of world reserves. As shown in Tables 10.4 and 10.5, Brazil holds a privileged position in oil reserves in the context of South and Central America (8.6 per cent of the world). The data contained in the tables have been compiled by BP plc experts, using a combination of primary official sources, third-party data from the OPEC Secretariat, World Oil, Oil & Gas Journal and independent information in the public domain. Reserves include gas condensate and natural gas liquids (NGLs) as well as crude oil.

Oil & Gas Cluster in Rio de Janeiro 329

Table 10.4

Oil proven reserves at end of 2006

Region

Thousand million barrels

Total North America Total South & Central America Total Europe & Eurasia Total Middle East Total Africa Total Asia Pacific

59.9 103.5 144.4 742.7 117.2 40.5

Source: Authors, based on the BP Statistical Review of World Energy (2007).

Table 10.5 Oil proven reserves at end of 2006 in South and Central America South & Central America Argentina Brazil Colombia Ecuador Peru Trinidad & Tobago Venezuela Other S. & Cent. America Total S. & Cent. America

Thousand million barrels Share 2.0 12.2 1.5 4.7 1.1 0.8 80.0 1.3 103.5

0.2% 1.0% 0.1% 0.4% 0.1% 0.1% 6.6% 0.1% 8.6%

Source: Authors, based on the The BP Statistical Review of World Energy (2007).

Brazil recently achieved self-sufficiency in oil, although it still imports high grade oil for its refineries. PETROBRAS produces about 1.8 million barrels of oil a day, satisfying the Brazilian demand (Figure 10.1). But the company still needs to import light oil to mix with the local product heaviest in the process of refining. According to the state national company PETROBRAS,8 with its emphasis on exploration and production, its proven reserves (SPE criteria9) will jump to 15,000 million barrels and production will reach about 2.2 bby by 2015 and 4 bby by 2030.10 Most reserves and production come from offshore fields, with the frontiers of deep and ultra-deep water and depth exploration increasing every year (Figure 10.2). PETROBRAS is a publicly listed company that operates on an integrated and specialized basis in the segments of the oil, gas and energy sector (exploration and production; refining, commercialization, transportation

A. J. J. Botelho and G. M. Bastos

2003

330

16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0

Production Consumption

Russia Figure 10.1

China

Brazil

India

Brazil oil production and consumption

Source: Oliveira (2006) (PPT).

Billions of BOED 12

11.6

11.8

11.7

11.5

11.5 11 10.5

10.5 10 9.5 2002

2003

2004

2005

2006

Figure 10.2 Brazil proven reserves of oil and NGL at end of 2006 (in BOE – Billion barrels of oil equivalent) Source: Authors, based on the PETROBRAS Annual Report 2006 (2006 Operational Figures).

and petrochemicals; distribution of oil products; natural gas and energy), was founded in 1953, and is now the world’s 14th largest oil company, in accordance to the publication Petroleum Intelligence Weekly.11 Despite the expressive presence of foreign companies such as RepsolYPF, Exxon Mobil, Chevron, Agip, etc., companies that operate heavily in Brazil in the exploration and production segment, in this chapter we actually give greater emphasis to PETROBRAS due its very effective role in the ‘induction process’ of the development of an industrial cluster in Campos Basin, considered the biggest oil reserve in the Brazilian Continental Platform. In a way, it would not be a big mistake to consider that the proto-cluster in Campos Basin was formed almost ‘tailored’ to meet its demand.

Oil & Gas Cluster in Rio de Janeiro 331

10.3 The state of Rio de Janeiro 10.3.1 The Campos Basin (off Rio de Janeiro state coast) Campos Basin, located mainly off the northern coast of the state of Rio de Janeiro, has the largest oil reserve in the Brazilian Continental Platform, with measures of some 100,000 square kilometres and ranges from the State of Espírito Santo, near the city of Vitória; to the city of Arraial do Cabo, off the northern coast of the state of Rio de Janeiro. Campos Basin is the largest oil province in Brazil, responsible for around 83 per cent of the country’s oil production and 77 per cent of PETROBRAS recoverable oil reserves, with a daily production of approximately 1.5 million barrels of oil and 22 million cubic metres of natural gas, 48 producing fields, 14 fixed platforms, one fixed re-pumping platform, and 25 floating production systems. Although the fields in the Campos Basin have been in production for quite some time (30 years), it is not presently a mature oil province (Formigli 2007). Much to the contrary, as Formigli puts it, the Campos Basin is ‘the largest open-air laboratory in the Brazilian offshore sector, where PETROBRAS’ most important technological innovations utilized in deep and ultra deepwater exploration and production were tested on a large scale’. Having started production in 1977, the Campos Basin is in the PETROBRAS Program for the Revitalization of Oil Fields with a High Degree of Exploitation, the most important programme in the exploration and production area; implements diversified techniques, used according to the necessity, to improve the recovery factor its fields. The targets are ambitious, as Formigli (2007)12 stated: In the Basin, the average extraction factor of oil in place is presently 27%. But we are working to raise this index to 40%. In the case of the Marlim field, the largest producer in the Campos Basin, we have already achieved the projected value of 44%. However, the target we hope for exceeds 50%. Thanks to improvements in the methodologies and technology of geological analysis, and principally to the evolution of seismic survey technology from the 1970s to date, it is now possible to find reservoirs not previously detected in the basin, more than 8,000 metres below the seabed (deep off-shore). Thus, recent discoveries of light oil in areas adjacent to the Campos Basin and in the Espírito Santo and Santos basins could be indications that there are new fields at greater depths in the region. Thus, in March 2007, PETROBRAS announced the discovery of reservoirs saturated with light oil of around 30º API, below a thick layer of salt in the marine subsoil to the north of the Campos Basin, at a depth of more than 4,000 metres. The results of drilling the sizing wells are yet to be announced.

332

A. J. J. Botelho and G. M. Bastos

There are also high expectations for new discoveries in the pre-salt section, the rock formation below the reservoirs presently in production. Further, in November 2007, PETROBRAS announced the discovery of a pre-salt ultra-deep giant field, in the adjacent Santos Basin (Tupi), with estimated reserves of 5 to 8 billion barrels of high quality oil, increasing the company’s reserves by up to 60 per cent. Tupi13 could produce up to a million barrels per day, over half of current production capacity, and around 33 million cubic feet of gas. Before the discovery, PETROBRAS had announced investments of US$112.4 billion between 2008 and 2012, to meet a target daily production of 3.5 million oil barrels and gas equivalents, an amount 27 per cent higher than that of the period 2003–2007 (Figure 10.3). The majority of the investments in the country (87 per cent) will be in exploration and production (US$49.3 billion). The state of Rio de Janeiro will receive US$40.5 billion (36 per cent) of the total investment. Over the period, taking into account a local content of the projects of 60 per cent, industrial purchases in Brazil will be of US$24 billion. In recent years, the economy of the state of Rio de Janeiro has become heavily dependent on the O&G industry. Royalties and other mandatory contributions received by the state jumped from R$1.43 billion in 2000 to R$4.14 billion in 2006, 60 per cent of all production royalties paid. In the state, the municipality of Rio de Janeiro has (2005) the country’s

1,800 US$1.53 billion/year

US$1.12 billion/year

1,600

1,464 1,400

US$880 billion/year 1,163

1,200 956

1,000 840

US$536 billion/year

909

1,030 833

814

800

725 549

600 400

505

566 448

533

571

393

200 0 1991

Figure 10.3

1994

1998

2004

PETROBRAS evolution investments, 1991–2001

Source: Authors, based on PETROBRAS (2007).

2006

2011

Oil & Gas Cluster in Rio de Janeiro 333

second largest GDP, less than half of the first placed, São Paulo (R$263.18 billion). Campos dos Goytacazes in the Campos Basin region was ranked fifth in 2002. 10.3.2 Regional innovation system The state of Rio de Janeiro has one of the largest and most developed regional systems of innovation14 in the country, after the state of São Paulo. It has several (eight out of 147 institutions of higher education) research universities (including four federal, led by the country’s largest federal research university UFRJ), the country’s largest private research university PUC Rio and a leading military engineering research university Instituto Militar de Engenharia (IME), over a dozen national research institutions and technological institutes (including the leading CBPF Centro Brasileiro de Pesquisas Físicas; CETEM – Centro de Tecnologia Mineral; INT Instituto Nacional de Tecnologia; LNCC Laboratório Nacional de Computação Científica and IMPA Instituto de Matemática Pura e Aplicada) and the largest number of technological incubators, with 133 incubated start-ups and having ‘graduated’ 102 start-ups, and seven local production systems, including the O&G industrial agglomeration of the Campos Basin. According to CNPq (The National Council for Scientific and Technological Development), in 2004 there were 7,597 PhDs in the state for resident population of 15.2 million, a rate of 50 PhDs per 100,000 inhabitants, and the largest rate of student registration and doctorates (next to the state of Rio Grande do Sul) among those states with the highest schooling ratios in the country. In the area of engineering, of the existing 205 graduate courses, 34 or almost 17 per cent are in the state. Twelve of these have the highest grades (6 and 7) in the evaluation of the Ministry of Education agency CAPES. The state also contains 46 per cent of the centres of excellence (ahead of São Paulo with 38 per cent). An important bridge institution of the system, particularly for the Rio de Janeiro metropolitan region local innovation system is the Rio de Janeiro Technology Network (Rede de Tecnologia do Rio de Janeiro: REDETEC) a civil not for profit organization aimed at the technological diffusion and support, through the matching of supply and demand and stimulation of public and private support agencies and other governments instruments and mechanisms. REDETEC is also the executive secretariat of the Network of Technological Incubators, Parks and Poles of the State of Rio de Janeiro (Rede de Incubadoras, Pólos e Parques Tecnológicos do Estado do Rio de Janeiro: ReINC).

10.4

Evolution of the national O&G supply chain

For over five decades of its existence, PETROBRAS had a significant role in the development of national industry provider of goods and services to oil

334

A. J. J. Botelho and G. M. Bastos

and gas activities. By the end of the 1980s the process of replacing imports conducted within the sector of oil and gas allowed a national Supply Chain Network to be built from raw materials to components and finished products. Based on successful programmes of training and qualification of local suppliers, conducted by PETROBRAS, the domestic industry has to be achieved, high levels of care to more than 90 per cent of direct purchases of materials placed on the country. According to ONIP,15 the model was victorious in enlarging the capacity of local supply, but little attention has been given to issues of training and technological innovation as well as modernization of business management. In other words, the industry has developed without due consideration for competitiveness in international standards. During a serious foreign exchange crisis of the 1980s, there predominated in Brazil a clear guidance of industrial policy inducing the economic protectionism and the purchase of domestic goods and services, even in non-efficient conditions for price, production capabilities and quality. The opening up of the Brazilian economy which occurred in the 1990s placed the national industry facing international competition, in most cases in unfavourable conditions on the financial costs and tax issues. Despite the effects suffered by the sudden exposure to international competition, the national Supply Chain Network was preserved in its essential aspects. Throughout the 1990s and up to the present day, the level of direct purchases of materials of PETROBRAS on the domestic market has always been above 75 per cent. Large enterprises of exploitation marked the priority investment of PETROBRAS from the 1990s, surpassing the ‘milestone’ of US$10 billion in the second half of the 1990s, and that most of these investments focused on the businesses in deep water. Yet according to ONIP, in this period – when there was no explicit guidance as to a concrete policy of national supply – the local content of goods and services in the acquisition of units floating production ranged from 35 to 52 per cent when the units were built in the country and from 1 to 19 per cent when they were ordered in from outside. ONIP is composed of almost 30 supporting members. These members represent National Associations from the different sub-sectors of the oil and gas industry, including: exploration, drilling, engineering, tubes and pipes, naval construction and design, heavy machinery manufacturing, infrastructure, electric/electronic materials, services and others. In addition to the main oil companies and operators already operating in Brazil, including PETROBRAS, three states governments (Espírito Santo, Rio de Janeiro and Rio Grande do Norte) as well as the federal government (Ministry of Development, Industry and Foreign Trade – MDIC) are also affiliated to ONIP. ONIP promotes the interaction of local suppliers with anchor firms,16 creates forums aiming to facilitate the achievement of partnerships between domestic suppliers and foreign buyers and try to help in

Oil & Gas Cluster in Rio de Janeiro 335

the elimination of barriers to the full development local industry of goods and services. The opening of the national oil industry took place in a time of depreciated oil prices in the international market. As the new national policy prioritized the attraction of foreign investment in the industry, focus on creating competitive conditions for foreign investment in the sector was in detriment of the national local industry. In recent years, several actions were led by ONIP, seeking, for instance, to tax adjustment policies for the domestic supplier. However, legal implications and non-acceptance on the part of tax authorities state (Federal Revenue Office) prevent the disposal of such instruments. On the other hand, in 2003, aiming to expand the participation of the national industry in O&G ventures, in a competitive and sustainable basis, the Brazilian federal government created the programme ‘Mobilization of Industry National Petroleum and Natural Gas (PROMINP)’, coordinated by the Ministry of Mines and Energy (MME) and PETROBRAS, in a joint effort to identify the demand for goods and services, mapping the local productive capacity to full development of a local supply chain. In order to implement a policy that resulted in an immediate increase of the level of local content and employment in the country, the Federal Administration introduced, in 2003, significant changes in the treatment of the issue of the national supply business in exploration and production (E&P) of oil and natural gas, consisting of new rules for the exploration area biddings of the National Petroleum Agency – ANP (Agência Nacional do Petróleo, Gás Natural e Biocombustiveis) and the requirements of local content in the biddings for PETROBRAS platform constructions. It is necessary to emphasize that ANP is a special autarchy created by Law 9,478, dated 6 August 1997, as an integral part of the Indirect Federal Administration, entailed to the Ministry of Mines and Energy (MME), and is the only instance that can announce and bid the opening of ‘Brazil Rounds’, in which offshore and onshore blocks, distributed among basins, are offered for exploratory purposes by anchor firms in the position of ‘Concessionaires’. Also, only ANP defines the requirements, procedures and calculation of local content per block in each ‘Brazil Round’. The ANP and Ministry of Mines and Energy (MME) goals with the Local Content Politics17 can be so described: ●

●

● ●

Increment of the participation of the national industry in competitive bases; Increment of the qualification and the national technological development; Increment of the professional qualification; Employment and income generation.

336 A. J. J. Botelho and G. M. Bastos

It should be stressed that the anchor firms in the position of ‘Concessionaires’ shall execute a Concession Agreement for the Exploration, Development and Production of Oil, Natural Gas for each Block within the Concession Area, as the ANP requires that in the Exploration Phase, ‘Concessionaires’ must purchase an amount of goods and services from Brazilian Suppliers so that the minimum Local Investment Percentage, respectively, is 70% (seventy percent) onshore, 51% (fifty-one percent) in shallow waters with depth equal or inferior to 100 meters and 37% (thirtyseven percent) in shallow waters with depth between 100 and 400 meters, as well as in deep waters (ANP 2007). The key role that the ANP performs in terms of induction into the process of technological innovation of suppliers involved in the biddings seems very clear, even with the existence, in the Concession Agreement for the Exploration, Development and Production of Oil and Natural Gas, of clauses that propose a local content to the Concessionaires. In general, the Concessionaires, in fulfilling the ANP Contracts, undertakes to: ● ●

●

●

include Brazilian Suppliers in the companies invited to submit proposals; ensure that all the invited companies shall have equal and adequate time consistent with the requirements of the Concessionaire, both in the preparation of proposals and in the delivery of goods and services, in accordance with the best practices of the oil industry, so as not to exclude potential Brazilian suppliers; require no technical qualifications or certifications of Brazilian suppliers besides those required from foreign suppliers; keep track of the Brazilian suppliers that are able to offer supplying services and seek, whenever applicable, updated information on the universe of suppliers at the trade associations and entities with renowned knowledge on the subject.

10.5

Innovation and capacity building in O&G sector

10.5.1 Innovation government policies 10.5.1.1

CT Petro

Following the liberalization of the oil exploration market in Brazil in 1997, which broke PETROBRAS’s monopoly in oil exploration and production, several government and national industrial association efforts were made to ensure the continuing technological development of PETROBRAS and of its national suppliers as a critical element in the company’s competitive growth. One of the first to be put in place was the Oil and Natural Gas Sector Fund (Plano Nacional de Ciência e Tecnologia

Oil & Gas Cluster in Rio de Janeiro 337

do Setor de Petróleo e Gás Natural, also known as Fundo Setorial para Industria de Petróleo e Gás: CT Petro) launched in 1998, the first in a series of funds that have contributed to a rise in the level and stability of STI financing, both for research institutions and universities as well as for firms. In addition, the Sector Funds established a new management and governance model based on the participation of various social segments within each sector; on the effort to establish long-term strategic goals and on the focus on results and definition of priorities. Resources for the Sector Funds usually come from small percentages of existing mandatory financial contributions (e.g., on exploitation of natural resources), taxes (e.g., Industrial Production Tax – IPI) and levies (e.g., on remittances for payments of technical assistance and technology transfer) on productive revenues in each sector (see Table 10.6). Today there are 16 Sector Funds (SF) in operation, including two so-called horizontal ones. Revenues for all funds, with the exception of the telecommunications SF – Funttel, are directed to the FNDCT, and are managed by its executive secretariat, FINEP, Brazil’s innovation agency, a public enterprise under the Ministry of Science and Technology (MCT). In order to promote a geographic de-concentration of research activities, it is mandatory that 40 per cent of CT Petro’s and 30 per cent of all other SFs’ disbursements must be made in the north, north-east and centre-west regions. A growing share of SFs resources are being managed and implemented by the CNPq through its grants and scholarship mechanisms. CT Petro is the largest among the vertical funds (and third overall in 2005), with an executed budget of R$75.6 million. Nowadays, only four firms contribute resources to CT Petro, but nearly all of it comes from PETROBRAS, the single largest producer which also submits the most R&D projects.18 At the end of 2005, cumulative revenues had reached about R$1 billion, but the CT Petro budget for funding R&D projects reached R$150 million (US$88 million) in 2006, as the majority of revenues in held in Treasury by the government to assist in the formation of budget superavit. CT Petro governance is made of a committee which sets priorities, based on studies and internal evaluations,19 allocates resources and supervises targeted public calls. Its major tasks are to assess collection of research activities funded and allocate resources among proposals, but it also evaluates directly large projects. The fund is operated and managed by FINEP. CT Petro funds joint university–industry projects in which the company funds 50 per cent and CT Petro the other 50 per cent. In 2006, CT Petro had four calls. The latter resources are divided between a programme for medium-sized and large companies and another for SMEs, including local production systems (Arranjos Produtivos Locais: APLs). Only eight APLs out of 300 identified by the quasi-public Micro and Small Enterprise Support

338

A. J. J. Botelho and G. M. Bastos

Table 10.6

Sectoral funds: Regulatory frame and resources

Fund

Regulatory frame

Resources

CT-Petro

30/11/1998

CT-Info

20/04/2001

CT-Infra CT-Energ CT-Mineral

26/04/2001 16/07/2001 16/07/2001

CT-Hidro

19/07/2001

CT-Espacial

12/09/2001

CT-Saúde CT-Biotec CT-Agro CT-Aero Fundo VerdeAmarelo

25/02/2002 07/03/2002 12/03/2002 02/04/2002 11/04/2002

CT-Transp

06/08/2002

CT-Amazônia

01/10/2002

25% of the royalties that exceed 5% of oil and natural gas production Minimum of 0.5% of the gross income of companies benefited by the ‘Lei de Informática’ 20% of the resources of each sectoral fund 0.75% to 1% gross income from concessions 2% of financial compensation (Cfem) payments by firms for access to mineral exploration 4% of the financial compensation used for electrical energy generators 25% of income used for orbital positions; total income from licences and authorizations from the Brazilian Space Agency 17.5% Cidea 7.5% from Cide 17.5% from Cide 7.5% from Cide 50% from Cide, 43% from the income from IPI stemming from products benefited by the ‘Lei de Informática’ 10% of the income from the National Transport Infrastructure Department (contracts for the use of ground transport infrastructure) Minimum of 0.5% of the gross income from computer firms in the Zona Franca de Manaus

Note: a Cide means Contribution of Intervention in the Economic Field incident on the import and marketing of gasoline and its currents, and their current diesel, fuel oils, liquefied petroleum gas (LPG), including the derivative of natural gas and naphtha. Source: Authors, based on Pereira (2005).

Agency (Serviço de Apoio a Micro e Pequena Empresa: SEBRAE) have been funded. CT Petro also funds supplier development projects in the framework of the Brazil Technology Network (Rede Brasileira de Tecnologia: RBT), upon request by PETROBRAS to CT Petro to launch a special public call, with proposals evaluated by PETROBRAS. In 2006, this programme line spent R$66 million on 16 calls (eight defined by PETROBRAS and eight by CT Petro).

Oil & Gas Cluster in Rio de Janeiro 339

The fund has priority strategic themes such as research on heavy oils and pipelines. For example, it funded the Pipeline Technology Centre (CTDUT) with PUC Rio, which is another link next to the Brazilian Pipeline Competence Network (RBCD in RJ) between ICTs and funding agencies and pipeline operating firms (PETROBRAS, Transpetro, TBG, TSB and others). More than half of the R$40 million invested in 2006 went to Rio de Janeiro ICTs, mainly universities. Finally, in 2006, some CT Petro resources also began to be invested in horizontal activities for R&D such as metrology and basic industrial technology. As stipulated in the regulatory law, 40 per cent of CT Petro expenditures go to ICTs in the north and north-east regions, about R$60 million in 2006. The majority of the free resources go to the largest research universities with accumulated expertise in the sector, primarily UFRJ and PUC, as well as Unicamp, USP, UFRN and UFBA. In Rio de Janeiro other universities such as UFF and UERJ (state university) also receive a considerable amount of CT Petro funds. Moreover all the big projects are at Rio de Janeiro universities, such as the oceanic tank (the world’s largest) at UFRJ and the above mentioned CTDUT at PUC Rio. Overall, about 25 per cent of CT Petro resources go to Rio de Janeiro ICTs. 10.5.1.2

Special Participation clause

Another major source of resources for R&D in the sector is a mandatory fiscal contribution, the so-called ‘Special Participation’ (Participação Especial: PE), made by firms exploring, developing and producing oil and gas as stipulated in the R&D clause of the concession contract signed with the sector regulatory agency National Petroleum Agency (Agência Nacional do Petróleo, Gás Natural e Biocombustiveis: ANP), which has as among its goals to induce the development of national technological solutions for the sector. This contractual obligation determines firms operating high profitability production or high volume production fields (thus obliged to pay an additional PE in royalties to local and regional authorities) must allocate 1 per cent of revenues for R&D projects, up to half for in-house and at least half for extra-mural research in accredited research institutions. The resources are managed by the regulatory agency ANP. At least 50 per cent are directed to fund R&D projects in ANP accredited science and technology institutional (Instituições de Ciência e Tecnologia: ICT, which comprises universities and research centres). Only in November 2005 ANP passed a norm to make this clause operational.20 Again, PETROBRAS is the main investor under this clause, although Shell has also begun to fund R&D projects under it. With the ANP authorization, through the Special Participation clause scheme, between 2005 and 2006, O&G production firms, mainly PETROBRAS, contributed about R$620 million (accumulated resources over the 1998–2004 period) for PE projects. In the first phase, from December 2005 to March 2006,

340 A. J. J. Botelho and G. M. Bastos

PETROBRAS invested R$158 million, in the professional capacity building programme (Programa de Mobilização da Indústria Nacional de Petróleo e Gás: Prominp), which has an objective to train 60,000 professionals, a number that will double by 2011 (please see below). In a second phase, from April 2006 to February 2007, it funded 208 R&D projects (out of 265 submitted) to the total of R$457.8 million, with the near majority of resources directed to laboratory infra-structure (99.2 per cent). Forty projects are still under evaluation (R$100.14 million) and 17 were cancelled or did not meet the criteria. PETROBRAS R&D management structure, already overwhelmed by ongoing cooperative research contracts with universities (120-plus), decided to adopt a different strategy here. It structured its investments along two main axes: (1) development of 33 Thematic Networks (Redes Temáticas) – in the areas of exploration, production, supply, gas, energy and sustainable development and management and innovation (see Figure 10.4); and (2) establishment of seven regional competence centres in partner research institutions. For example the heavy oil thematic network comprises five universities and one research centre. Regional competence centres are also networkbased, but facilities are built in the lead institution.

427.6 25% Regional centres

121.8

1% Others

74% Thematic networks

16.8

Thematic networks Figure 10.4

Regional Others centres

PETROBRAS technological cooperation programmes

Source: Authors, based on PETROBRAS (2007).

Oil & Gas Cluster in Rio de Janeiro 341

In addition, requests for accreditation to ANP and proposals to PETROBRAS had to fall under the following Technological Services Groups: 1. Development and engineering of operational units and commodities inputs; 2. Development of product and engineering processes; 3. Development of information systems and software to control or processing; 4. Development of products and processes for monitoring, management and conservation of the environment; 5. Development of methodologies for laboratory tests; 6. Project management programmes; 7. Human resources development. So far out of the 50 universities which have received a total of R$488 million from these funds, 28 are located in the south-east region. The two main research universities in the state of Rio de Janeiro, PUC Rio and Federal University of Rio de Janeiro (UFRJ) have received R$50 million and R$105 million respectively (see Figure 10.5). Overall, Rio de Janeiro ICTs received about two-thirds of the funds. ANP authorizes PE activities as technological projects and/or programmes in the following categories: basic and applied research, experimental development (including pilot unit and prototype) and supplier development (including pilot manufacturing and industrial development projects) in the 80 69.5 70

40

Minas Gerais

50

Espirito Santo

60

Rio de Janeiro

São Paulo

45.5

24.6

30 20

18.0

16.3

19.0

6.7

10

CT-Graphics UFES UFJF UFMG UFU UNIFEI CEFET Campos CETEMMCT CTDUT-RJ CTEx FBTS IME INMETRO INT Marinha ONMCT PUC-Rio UENF UERJ UFF URFJ UFRRJ ABTLUS INPE IPT-SP ITA UFSCar UNESP UNESP R C UNICAMP USP

0

Figure 10.5

Regional distribution of PETROBRAS cooperation

Source: Authors, based on PETROBRAS in Oliveira (2007).

342

A. J. J. Botelho and G. M. Bastos

areas oil and its derivatives, natural gas, energy, environment and human resources training. 10.5.1.3

Programme for human resources training for the O&G sector (PRH)

The PRH of ANP congregates a total 44 university-based programmes in 31 higher education institutions in 16 states, of which 23 are universities. These institutions in articulation with industry effectively co-manage the programme. It has made expenditures of R$103 million in training human resources in all knowledge areas connected to the oil sector, improving teaching and research infrastructure and providing scholarships to over 4,000 technical and higher education students in over 200 specializations of 44 training programmes. The programme is funded mainly with royalties’ resources of CT Petro. 10.5.2 PETROBRAS PETROBRAS made its first offshore discovery in 1974 and from 1981 began exploring and producing oil and gas from deep sea giant fields (Albacora 1984 and Marlin 1985; 300 m to 2,100 m), which present a high technological complexity and thus a very high risk, but conversely a high premium. PETROBRAS took the risk and in the past two decades developed and integrated the necessary technology from various sources to meet the growing challenge. Its strategy at the time was to launch a massive, second generation, capacity building programme with the goals of training 1,300 geologists to generate an understanding of the nature and process of O&G formation in and geological features and tectonic structures of the continental shelf (e.g., new models for turbidity formations); and 10,000 engineers to acquire domestic technological capabilities through learning by doing. In parallel, its research centre CENPES began to structure a large cooperative programme with mainly local universities to develop long range technological capabilities and critical expertise. This technological strategy was divided into two main overlapping phases. A first phase of incremental innovation (which lasted roughly until 1994) included the acquisition and assimilation of technologies (1960/1980), adaptations and absorption of design (1980/1990), joint R&D and knowledge production (1990/2000) and technology exchange and specialized knowledge production from 2000 (Oliveira 2007). An indicator of PETROBRAS growing technological capability was the fact that in 1992 it received the Distinguished Achievement Award (OTC). As E&P demands evolved so did PETROBRAS’s technological strategy, as shown in Figure 10.6. In the 1980s, pressed by the need to put into production the first shallow water offshore fields, the model was to acquire foreign technology. Thus the first seven fixed platforms in these fields followed the North Sea model, with the basic project acquired from abroad and detailed engineering done by Brazilian companies.

Oil & Gas Cluster in Rio de Janeiro 343

1977 1979 Enchova Bonito EN-1-RJS 1983 Piraúna RJS-38 124 m 1985 RJS-232 189 m Marimbá 293 m RJS-284 1988 383 m Marimbá 1992 RJS-3760 Marlin 492 m

1994 MRL-9 Marlin 1997 781 m MRL-4 2003 2000 1999 Marlin Sul Rencador Rencador Rencador MRL-3 1,027 RO-21 RO-8 RJS-436 1,709 m 1,886 m 1,877 m 1,853 m

Figure 10.6

PETROBRAS deep sea drilling technology evolution

Source: Authors, based on PETROBRAS in Oliveira (2007).

The last stages of this previous phase prepared the basis for the launching of the next, oriented to radical innovation with the capacity building programmes (Programa de Capacitação PETROBRAS) Procap 1000 (launched in 1984, aimed at O&G exploration at 1,000 metres water depth) and Procap 2000 and CENPES R&D in a radically new platform design which achieved a 30 per cent cost reduction. These programmes resulted from a decision by PETROBRAS based on the fact that at the time there were no proven deep sea production technologies available in the market to put into production the fields of Albacora and Marlin. The results of these technological efforts, such as wet Christmas trees (200 m); flexible risers; robotics (over 300 m); new anchors (plastics); new technologies for liquid flow; horizontal holes; drilling mud; resistance of drilling equipment; ship conversion to dual exploration and production platform; platform automation) provided for the discoveries of deep and ultra-deep oil fields and the timely launch of their production in the past decade, as shown in Figure 10.10. Again in 2001 and 2007 PETROBRAS was awarded the OTC prize. Furthermore, PETROBRAS holds the record of 1,870 metres of offshore oil exploration and a stud showed that for each dollar invested in R&D it had a return of US$8 dollars. Between 1999 and 2001, PETROBRAS and CT Petro funded R$13 million (US$7.6 million) for R&D in deep sea water exploration, involving

344

A. J. J. Botelho and G. M. Bastos

Strategic objectives

Technological prospecting

Technological demands

Competitive intelligence

Technological guidelines

R&D projects portfolio

Suppliers Figure 10.7

Internal development

R&D institutions

PETROBRAS technological system

Source: Authors, based on UFU (2007).

universities (UFRJ-Coppe, PUC Rio, USP, Unicamp, Universidade Federal do Rio Grande do Sul and the National Technology Institute (INT)). PETROBRAS investments in external R&D have been growing rapidly in the past few years, reaching about US$200 million in 2006, besides investments in its own R&D centre (Centro de Pesquisas Leopoldo Miguez: CENPES, please see below) and over 80 ICTs (mainly universities, including 11 in the state of Rio de Janeiro) through 38 thematic networks on O&G objectives. Results are diffused to PETROBRAS and its suppliers. UFRJ and PUC Rio are among the five largest recipients in Rio and account for about 65 per cent of the US$200 million. Overall, including PE and CT Petro resources in addition to PETROBRAS own budget R&D expenditures on universities jumped from R$40 million (US$24 million) in 2001 to about R$400 million (US$253 million) in 2006. Its technological system is depicted in Figure 10.7. The centrepiece of PETROBRAS’s strategy is its R&D centre CENPES, founded in 1974. PETROBRAS’s overall R&D expenditures have also been growing rapidly over the past few years. CENPES’s budget reached US$900 million (operational costs and investments) in 2006, a 25 per cent growth over the previous five years, or 0.9 per cent of revenues. CENPES employs 1,600 R&D personnel, including 700 full-time researchers, most with PhDs, and 700 technicians. CENPES accounts for 70 per cent of PETROBRAS’s overall R&D budget, with the remainder going into

Oil & Gas Cluster in Rio de Janeiro 345 R$ Million (Base 2005) 1,600 R&D expenditures 1,400 1,200 1,000 800 600 400 200 0 2001 Figure 10.8

2002

2003

2004

2005

2006

PETROBRAS R&D expenditures

Source: Authors, based on UFU (2007).

headquarters’ large projects. PETROBRAS is the organization with the largest number of deposited patents in the country in 2005 and 2006 and has a portfolio of over 1,000 patents. However, in order to meet its production target, which will grow by 7.8 per cent per year in the period 2007–2011 (Figure 10.9), PETROBRAS will have to even more quickly increase its R&D expenditures and, more importantly, complement its technological strategy with an innovation and largescale technological diffusion strategy. 10.5.3 Trends and perspectives In 2007, according to ANP, expenditures under the PE (Special Participation Clause) to finance ICT R&D projects will exceed R$670 million (US$394 million). Recently, in August 2007, PETROBRAS’s president, José Sergio Gabrielli, participated in platform P-54’s baptism ceremony. The FPSO-type21 unit is the second of its type built by PETROBRAS. The first was the P-52, which was baptized in June 2007 and is currently in its final testing phase to depart to the Campos Basin. The P-52 and the P-54 are part of the development programme for the ‘Roncador Field’, in the Campos Basin (state of Rio de Janeiro), and when they reach peak production, they will add 360,000 barrels per day to PETROBRAS’s production.

346

A. J. J. Botelho and G. M. Bastos 7.8% p.a. Oil and natural gas (Thousand boed)

8.7% p.a. 4,556

2,036

2,020

85 161 250

94 168 265

2,217

2,298

96 163

101 142

3,493

274

277

185 383

278 742 724

551 1,540 1,493

1,684

2,812

1,778

2,374

2003 2004

Target 2011

2005 2006

Figure 10.9

Forecast 20015

Oil and NGL - Brazil

Oil and NGL - International

Natural gas - Brazil

Natural gas - International

PETROBRAS production targets

Source: Authors, based on PETROBRAS (2007).

In order to meet the emerging technological challenges, CENPES, which currently has 1,865 employees, of which 1,046 have higher education diplomas (42 per cent with a Master’s and 16 per cent with a Doctoral degree) has a goal of having 80 per cent of its university level personnel with a Master’s or a Doctorate. PETROBRAS invested R$1.42 billion in R&D in 2006 and in 2007 close to R$1.8 billion (a little over US$1 billion at the year end exchange rate US$1 = R$1.7) and plans to invest R$2 billion (US$1.2 billion), reaching 1,500 research contracts with ICTs over the past five years. 10.5.4

O&G capacity building national policies

The major capacity building programme for the O&G sector is Prominp.22 The PROMINP, designed under the Ministry of Mines and Energy (MME), aims to strengthen the national industry of goods and services and is focused in the area of oil and natural gas. The goals of the programme, developed together with the industry, will lead to maximization of the participation of the national industry in the provision of goods and services,

Oil & Gas Cluster in Rio de Janeiro 347 R$ Million 50 45 40 35 30 25 20 15 10 5 0 2001

2002

2003

2004

2005

2006

2007

2008

Figure 10.10 Projected special participation expenditures on R&D Source: Authors, based on UFU (2007).

on a competitive and sustainable basis, given domestic and international demands, to generate employment and revenue income in the country, and to add value in the local supply chain. The PROMINP starts its activities now with a significant portfolio of 47 projects approved by the board of the programme (see Figure 10.11), in which are represented the government, businesses and entities in the class who serve in these activities. The challenge is to develop projects to increase the national or local content in specific areas of exploration and production, maritime transport, supply and gas and energy. Thus, the industry will be gradually improving itself to meet the demands, in the order of US$41 billion, from the investments that are being made in the sectors of oil and gas in the period from 2003 to 2007. 10.5.5 Innovation and capacity building in the cluster of the Campos Basin Trying to promote an insertion of small suppliers in the productive supply chain of oil and natural gas of the Campos Basin, and considering a scenario of potential and difficulties, PETROBRAS and SEBRAE23 signed in October 2004 an important national agreement, aiming to get (1) commitment of large anchor firms in the project, (2) diagnosis to identify opportunities for

348

A. J. J. Botelho and G. M. Bastos Ministry of Mines and Energy MME – Minister Ministry of Development, Industry and Foreign Trade MDIC– Minister PETROBRAS–CEO and Director of Services The Brazilian Development Bank (BNDES) – CEO Brazilian Institute of Oil, Gas and Biofuels (IBP) – CEO ONIP – General Director

MME Board

MME – Secretary MDIC – Secretary PETROBRAS – Executive Manager of Engineering PROMINP – Executive Coordinator ONIP – Director Associations – CEOs (CNI, ABCE, ABDIB, ABEMI, ABIMAQ, ABINEE, ABITAM and SINAVAL) Brazilian Institute of Oil, Gas and Biofuels (IBP) – Director BNDES – Director

Secretary of oil, natural gas and renewable fuels – MME Executive Committee Executive Coordinator

Sectoral Committees E&P

TM

GE&TD

O&G Industry

ABAST

Exploração e Produção/ Exploration and Production (E&P) Transporte Marítimo/Naval Transport (TM) GE&TD Gás e Energia e Transporte Dutoviário/Gas and Energy and Pipeline Transport. (GE&TD) Abastecimento/Distribution (ABAST)

Figure 10.11 Management structure of the programme (PROMINP) Source: Authors, based on ALMEIDA J. R. and LISBOA. V. – Executive Presentation in Forum – Status of Manpower and Suppliers in the City of Volta Redonda. Volta Redonda: PROMINP (2007).

small suppliers, (3) structuring of a corporate and institutional network of cooperation, named ‘Rede Petro BC’ and (4) effective training to improve the innovation potential of small firms in the cluster. According to Onoe et al. (2007), simultaneously with this mobilization of national coverage, PETROBRAS, local government, SEBRAE, FIRJAN (Federation of Industries of the State of Rio de Janeiro) and ONIP already triggered a link between the main actors in the chain of suppliers of the industry of oil and gas companies as suppliers of goods and services, including the operators – PETROBRAS in particular – aiming to develop practices for collaboration in support of the competitiveness of the Campos Basin cluster, building a trajectory in which relations among the players could be guided by a confidence and competitive environment. In 2003 there emerged the ‘Network’ named REDE PETRO BC (see Figure 10.12). According to Nader, G. (2007) REDE PETRO BC is helping PETROBRAS to increase the outsourcing of products and services from third parties, including activities of control and integration of the business and financial viability of projects, by changing its policy of acquisitions and contracting since 2004, in a clear result of the efficiency of the network and the effort of entities as ONIP and SEBRAE, with ongoing efforts for training, development, innovation in goods and services offered in the light of the

Oil & Gas Cluster in Rio de Janeiro 349

Executive committee

Technical committee

Development support team

Executive secretary

Board of members Figure 10.12

Organizational structure of REDE PETRO BC

Source: Authors, based on Nader, G. presentation (2007).

agreement between PETROBRAS and SEBRAE, under a methodology placed in 2003, and used until the present day, called GEOR,24 that means ‘the Result-Oriented Strategic Management’. GEOR is directing all structuring projects that will occur in the Campos Basin cluster, to follow the hierarchy (1) aiming at beneficiaries; (2) focusing on results; (3) condensing the strategic vision; and (4) ensuring commitment, timeliness and proximity vis-à-vis managerial action. The results obtained by REDE PETRO BC are pretty concrete since then, such as: ●

●

●

signature of an agreement between UENF/PETROBRAS/FUNDENOR for development of three new technologies identified between the demands disclosed by PETROBRAS in the Network; certification of Sensor Permanent Fund Pit (PDG – Pressure Downhole Gauge) by the Centre for Research of PETROBRAS (CENPES), developed and nationalized by the company Transcontrol25 (a firm from Rio de Janeiro with a subsidiary in the Campos Basin cluster); about US$50 million in business deals closed at The sixth Rio Oil & Gas 2006 Business Round, in goods and service orders from 200 small and mid-size Brazilian companies (most of them from the Campos Basin cluster).

Since the REDE PETRO BC establishment, firms able to provide for PETROBRAS increased by almost 11 per cent and companies in a position to participate through PETRONECT (PETROBRAS’s portal of e-business), reached nearly 7 per cent. Looking carefully at the studies of Silvestre and Dalcol (2006) and Onoe et al. (2007), and on the basis of the references mentioned so far, we can almost propose a basic standard, or a protopattern of cluster based on the Campos Basin focused in regional economic growth.

350

A. J. J. Botelho and G. M. Bastos

10.6 Conclusions PETROBRAS’s 2020 Strategic Plan of August 2007 has among its orientations: ‘Increase production and reserves of oil and gas, in a sustainable manner, and be recognized for excellence in performance in E &P Technology. Become a world reference in technologies that contribute to sustainable growth of the Company in industries of petroleum, natural gas, petrochemicals and biofuels.’ Its 2008–2012 business plan sets the following production targets for O&G productions: 3,058,000 boed (out of a total production target of 3,494,000) in 2012 and 3,455,000 boed in 2015 (of a total of 4,153,000 boed). It projects total investments of US$112.4 billion, until 2012 (an increase of almost 30 per cent in relation to the previous plan), an average of US$22.5 billion per year, being 87 per cent (US$97.4 billion) in Brazil. The third largest increase is in E&P (32 per cent) and new projects account for US$13.3 billion, US$10.9 billion due to cost increases in the heated services and equipment market, US$4.2 billion due to currency appreciation and the remainder due to factors such as change in project scope, business model, etc. The largest increase (39 per cent) is in corporate, which includes increases in R&D. Already, these investment projections and their respective R&D efforts will have to be revised due to the need to face the challenges of bringing into production the new ultra-deep oilfields (such as Tupi with estimated reserves between 5 and 8 bnb) in the pre-salt layer, in which eight of 15 wells drilled indicated oil, constituting an oil province of 800 kilometres in length by 200 km in width at depths between 4,500 and 7,000 metres. Today PETROBRAS produces at a maximum depth of 2,700 metres. Currently, the cost of developing a field in smaller depths to produce 150,000 bpd is US$1.2 billion. But as the depth doubles, the cost triples, according to a PETROBRAS expert. The technological effort is to reduce the production costs in these ultra-deep waters. The above analysis seems to confirm some of the lessons identified in a recent review of clusters, networks and innovation (Breschi and Malerba 2007, p. 23) that the accumulation of capabilities in key large actors is critical to the process of cluster formation and initial growth; and provide support to another of the review’s indication that labour market characteristics and spin-offs are major elements in cluster development. In addition, the review points out the need to ‘disentangle as much as possible the notions of proximity an spillovers and move beyond them to deeper concepts such as face-to-face contacts, social networks, and labor mobility ... more analytical and empirical attention should be devoted to the study of some key local institutions, especially to the workings of the local labor market ad to the set of rules governing the relations among employers and employees.’ A recent detailed micro-study of innovation in the Campos Basin cluster (Silvestre and Dalcol 2006) concluded that the geographic agglomeration

Oil & Gas Cluster in Rio de Janeiro 351

had a positive impact on innovative activities by some firms in the industrial agglomeration. It noted that there are important differences in attitudes and behaviour towards innovation according to the four different groups of firms present: (1) O&G operating firms (goods and services demand); (2) high technological complexity offshore goods and service providers (about 50 firms); (3) moderate technological complexity offshore goods and service providers; and (4) low technological complexity offshore goods and service providers. The study revealed that industrial agglomeration firm’s knowledge connections with organizations outside the agglomeration were more frequent (92 per cent of 25 valid events mapped in ten firms in groups 1 and 2, including PETROBRAS). More importantly, it showed that intraagglomeration knowledge connections were more important for goods suppliers in group 1 than for service providers in the same group, which can be explained by the system-effects of the demand which require a greater degree of integration from suppliers and thus greater number of interactions and innovation pathways. However, external knowledge connections relations were present in all 12 events of service providers, mainly with other corporate divisions and foreign R&D centres. The study suggests that geographically specific oil provinces mould patterns of technological specialization in life cycles. Thus, it suggests that policy should encourage the approximation of lower technological complexity national suppliers with foreign large complexity suppliers. Alternatively, and in addition to the above policy suggestion, we propose the amplification in scope and the deepening of the structure of the development of the local innovation system based on intense cooperation with universities, which has already produced a handful of successful spinoffs – for example, Pipeway, Gavea Sensors and Activa at PUC Rio alone. PETROBRAS’s inward-looking, centralized technological R&D strategy and limited innovation-oriented system has to expand and evolve towards an outward-looking open-innovation system. This system requires a more proactive support to the development of technological absorption and incremental innovation capabilities, and corresponding incentives and rewards, of national second-tier goods and service suppliers and, more importantly given the increase in radical innovation demands required by the next E&P frontier of the ultra-deep mega fields, a concerted and trustbased collaborative arrangement of generation of radical technologies and their entrepreneurial appropriation by spin-offs produced by partner local universities. PETROBRAS and cluster partner institutions have been moving too slowly in this direction. For example, PETROBRAS has yet to establish a corporate venture capital fund announced in early 2007 and has to license more of its shelf IP to entrepreneurs to develop them into innovations. Partner institutions’ main policy, the O&G Productive Chain programme has to take into account that the four regional agglomerations it

352 A. J. J. Botelho and G. M. Bastos

supports – Niterói (offshore and naval industry), Macaé e Campos (E&P), Duque de Caxias (petrochemical pole and refinery), and Rio de Janeiro have different innovation capacities and logics, requiring differentiated and targeted policies. This chapter specified a set of new conditions and elements for a flowchart approach to resource-based industry upgrading, expanding the applicability of the model in sector and development stage terms (Figure 10.13). The scope and scale of the upgrading challenge will change as new O&G players start or increase production in the area and began to share influence on the technological and strategic choices of both domestic and international suppliers and to also shape the industrial cluster to a multi-hierarchical format. Campos Basin’s unique characteristics notwithstanding, technologies developed for E&P here will eventually generate new technological capabilities and stimulate technological learning in

Step 1 – Agglomeration

Exploration and production (E&P) segment 1. 2. 3. 4.

A

Engineering Civil construction Construction & assembly Maintenance of operation

Federal administration (National Petroleum Agency – ANP) – (I)

Anchor firms and ‘Concessionaires’ – (II)

National associations from the different subsectors of the oil and gas industry – (III)

Strategic issues 1. Training (capacity building) 2. Instruments of industrial policy 3. Business performance of the suppliers

Agglomeration of exploration and production (E&P) segment`suppliers – Outside and local content Figure 10.13 Continued

B

Oil & Gas Cluster in Rio de Janeiro 353 Figure 10.13 Continued

(I), (II) and (III)

Step 2 – Innovation

Innovation and capacity-building 1. 2. 3. 4.

Step 3 – Region economic growth

B

Innovation gov. policies Special participation clause Human resource training PETROBRAS (Major anchor firm) R&D expenditures

Universities, R&D institutes and related firms in the metropolitan area of Rio de Janeiro 1. Innovation gov. policies 2. Technical qualifications and certifications support 3. Results of PETROBRAS (Major anchor firm) R&D expenditures 4. Entrepreneurs incubators focused in E&P segment 5. Technological park systems 6. Start-ups suppliers

Small suppliers in Campos Basin cluster C

C

A Step 4 – Business deals in E&P segment

1. Innovation gov. policies 2. Increment of local content, provided by small suppliers 3. Corporate and institutional network of cooperation (‘Rede Petro BC’) 4. Employment and income generation

Figure 10.13 A flowchart of the exploration and production (E&P) Segment of the O&G sector in the state of Rio de Janeiro Source: Botelho and Bastos, this volume (2008).

suppliers providing goods and services in other O&G basins with similar geological features (e.g., Angola) and/or oceanic environmental conditions (e.g., North Sea deep-sea exploration) where the modified flowchart model can assist local and national authorities in building innovation-oriented clusters.

354

A. J. J. Botelho and G. M. Bastos

As this expansion happens, PETROBRAS will be hard-pressed to maintain its organizational assets, mainly its human resources which are critical in the E&P segment. The company has already begun to advance responses to this challenge due to a different menace in the end of the 1990s, the exodus of professional talents due to the lack of high posts, and corresponding salaries at the top of the company’s hierarchical pyramid; and a problem appearing the horizon, a skewed age distribution in which 40 per cent of the workforce has an average age of 30 years and another 40 per cent has an average age of 50, nearing retirement due to the ‘good’ public function retirement rules. The company’s response was to implement a ‘Y’ career in which experienced managers and technical professionals can apply to a different consulting career track, which procures them higher salaries and greater functional mobility outside their traditional career track. Another major human resource problem facing PETROBRAS that can put strains on the cluster evolution is the fact that PETROBRAS will have to hire a huge amount of professionals over the next few years and therefore will up talent and push salaries up, making it harder for cluster MSMEs to hire high-calibre professionals. As PETROBRAS hires the cream of the crop, suppliers will have to make do with the less talented professional, which could also hurt technological learning and innovation in the cluster, and ultimately the very performance of PETROBRAS E&P activities.

Notes 1. Since the REDE PETRO BC establishment in October 2003, the number of small suppliers capable of providing goods and services to PETROBRAS, increased by almost 11 per cent and the share of MSMs registered to take part in PETROBRAS’s e-business reached nearly 7 per cent. 2. In particular, in Program 0419 – Development of Micro, Small and Mediumsized Enterprises and in the establishment of the Permanent Working Group for Arranjos Produtivos Locais (Brazil’s policy concept for proto-clusters) (GTP APL) by the Executive Order No 200, 03/08/04, and reissued on 24/10/2005 on 31/10/2006, composed of 33 government institutions and non-governmental institutions of national scope. 3. Comprises the municipalities of Belford Roxo, Duque de Caxias, Niterói, Petrópolis, Nova Iguaçu, Rio de Janeiro, São Gonçalo and São João de Meriti. 4. Comprises the municipalities of Armação de Búzios, Campos dos Goytacazes, Cabo Frio, Macaé, Rio das Ostras and São Pedro d’Aldeia. 5. Portuguese term ‘APL’ is generally used to refer to ‘local production system’ and ‘local production agglomeration’. 6. Statistics published in The BP Statistical Review are taken from government sources and published data. According to BP plc, country groupings are made purely for statistical purposes and are not intended to imply any judgement about political or economic standings.

Oil & Gas Cluster in Rio de Janeiro 355 7. Proven reserves of oil are generally taken to be those quantities that geological and engineering information indicates with reasonable certainty can be recovered in the future from known reservoirs under existing economic and operating conditions. 8. PETROBRAS Annual Report 2006. 9. Society of Petroleum Engineers (www.spe.org/). 10. The figure includes information from reserves held abroad, corresponding to PETROBRAS’s participation in the partnerships, with reserves measured in accordance with the criteria of the Securities and Exchange Commission (SEC). 11. www.piwpubs.com/. 12. PETROBRAS Magazine, edition 52 (2007). 13. PETROBRAS’s partners in the Tupi oilfield include Britain’s BG Group, Galp of Portugal and Spain’s Repsol. 14. This section is based on Redetec (2007). 15. The National Organization of the Petroleum Industry (ONIP) is a Brazilian private and non-profit organization aiming to maximize the benefits resulting from the expansion cycle of the oil & gas sector for the Brazilian society. More details on www.onip.org.br. 16. Among anchor firms that are ONIP partners, we can cite companies such as Texaco, Devon, Exxon-Mobil, Maersk, Repsol-YPF, Shell and Total Fina (Total S.A.). 17. The rules for Local Content Certification can be accessed at www.brasil-rounds. gov.br. 18. In 2006 there was just one project from another firm, the O&G industry exploration and development service provider Halliburton. 19. For example, in 2001 under the coordination of the ANP, CT Petro contracted with the National Institute of Technology (Instituto Nacional de Tecnologia: INT) a study of technological trends in order to subsidize the decision-making process. 20. ANP’s resolution that approves the regulation for accomplishment of the investments in research and development and regulation of consolidated statement of financial accounting standards reports over the carried through expenditures. 21. Floating Production, Storage and Offloading Vessel. 22. www.prominp.com.br. 23. SEBRAE (the Brazilian Service of Support for Micro and Small Enterprises) was founded in 1972 and since then has been operating with funds from private initiative and from a compulsory contribution of 0.3 per cent and 0.6 per cent calculated on the total the payroll of Brazilian companies (chapter 8, paragraph 3 of the Law number 8,029, which created SEBRAE). SEBRAE is well known in Brazil by its strong commitment to the development of micro and small enterprises in the country. 24. www.sigeor.sebrae.com.br/. 25. www.transcontrol.com.br.

References ANP – Agência Nacional do Petróleo, Gás Natural e Biocombustíveis (2007) Resolução de Diretoria n° 655, de novembro de 2007, referente á Cláusula de Conteúdo Local constante dos Contratos de Concessão para Exploração, Desenvolvimento e Produção

356

A. J. J. Botelho and G. M. Bastos

de Petróleo e/ou Gás Natural, estabelecidos entre a ANP e os concessionários a partir de 2005. Breschi, S. and F. Malerba (2007) Clusters, Networks and Innovation. Oxford: Oxford University Press, p. 499. Britto, J. (2007) Mapeamento de Arranjos Produtivos no Estado do Rio de Janeiro. Executive Presentation. REDE SIST – BNDES, Rio de Janeiro, Abril. Formigli, J. (2007) Pre-Salt Reservoirs Offshore Brazil: Perspectives and Challenges. Miami: Energy Conference, November. Nader, G. (2007) REDE PETRO BC – Executive Presentation (A Rede Petro BC). Rio de Janeiro: SEBRAE-RJ. Oliveira, A. (2007) Hydrocarbons Challenges and Opportunities. PowerPoint presentation to Energy and Environment Innovation Systems Section. BRICS Conference, Rio de Janeiro, 25–27 April. Onoe, E., G. Nader, A. Batista, and R. Regazzi (2007) Projeto de APL de Petróleo e Gás da Bacia de Campos com Adequação ao Programa de Desenvolvimento de Clusters Industriais. 1. ed. Rio de Janeiro: SEBRAE-RJ, V. 1. p. 72. PETROBRAS Magazine, edition 52 (2007). PETROBRAS Annual Report (2006). REDETEC (2007) – Executive Presentation (Potencial do Rio de Janeiro). Silvestre, B.S. and P.R.T. Dalcol (2006). Abordagens de Clusters e de Sistemas de Inovação: Um modelo híbrido para análise de aglomerações industriais tecnologicamente dinâmicas. Revista Gestão Industrial (Online), Vol. 2, pp. 99–111. UFU (2007) Financiamento Petrobras. Universidade Federal de Uberaba, Pró-Reitoria de Pesquisa e Pós-Graduação, Agência Intelecto. PowerPoint presentation, 46 slides. www2.PETROBRAS.com.br (accessed 9 January 2008). www.spe.org (accessed 9 January 2008).

11 Conclusion Akifumi Kuchiki and Masatsugu Tsuji

This book provided a critical useful framework to explain the formation of agglomeration and the endogenous innovation process and to construct models to upgrade industrial clusters to knowledge-based ones. The underlying theory of these processes is ‘the Flowchart Approach’, which we have been proposing in Kuchiki and Tsuji (2005, 2008), and Tsuji et al. (2007). The objective of this book is to formulate the transforming process from agglomeration to endogenous innovation by expanding the Flowchart Approach. As a conclusion, this process is divided into two sub-processes, namely, agglomeration and innovation, and we postulated these processes according to the Flowchart Approach. It is referred to as Step I of agglomeration and Step II of innovation. This process cannot be a single path common to all economies, developed or developing, but surely there are various routes to achieve goals. In order to establish hypotheses, therefore, this book viewed various cases of clusters which are different in stages of economic development, in stages of agglomeration, and in industrial structures. Cases this book examined include small- and medium-scaled industrial clusters in China, software clusters in India; Japan’s SMEs (small and medium-sized enterprises) clusters, automobile clusters in Malaysia, industrial clusters and firm-provided training evidences in Bangkok, Thailand, biotechnology clusters in Singapore, and software development and the oil/gas cluster in Rio de Janeiro, Brazil. By analysing the cases mentioned above, the following points are extracted as important insights in transforming industrial clusters to innovative ones. First, the industrial cluster policy, which is widely analysed and proposed in our previous books, is still a key to this transformation. We proposed an industrial cluster policy based on the Flowchart Approach for aiming innovation. For this policy, the following points are crucial: First, local governments play an important role in the industrial cluster policy. The difference in policy for agglomeration and innovation lies in necessity of ‘the Local Innovation System’. Almost all economies concerned with innovation already established ‘the National Innovation System (NIS)’. For innovation 357

358 Akifumi Kuchiki and Masatsugu Tsuji

clusters, it is imperative to construct networks of all local entities related to innovation such as universities, business organizations, R&D institutions, local governments or MNCs (multinational corporations). Secondly, the industrial cluster policy for innovation is different in a targeting industry; that is, models based on the Flowchart Approach to the manufacturing industry cluster are different from those to biotechnology, for example, since innovation in biotechnology requires much more advanced human resources as well as higher technologies than manufacturing. Thirdly, innovation in larger firms is different to that of SMEs. There are fundamental difference in theory and policy between the agglomeration and innovation, but the Flowchart Approach we proposed for agglomeration can be theoretically and practically useful for transforming to an innovation cluster. By focusing on agglomeration and innovation issues, which are central and yet less investigated fields of economic geography due to lack of information on global and local knowledge linkages, this book contributes to the understanding of the process of transforming to innovation clusters from industrial clusters. Each chapter was prepared with in-depth interviews, field works, and micro-econometrics with careful treatment of global and local linkages. While approaches may vary, an overall connection on the clustering-innovation debate is maintained. Exposed to the global competition, firms seek for better business environments. Location selections by the private sector, especially multinationals, influence a great deal of industrial developments in developing countries. In these circumstances, local governments are required to take into consideration factors that encourage the agglomeration of firms in particular regions, and to design and implement effectively the industrial cluster policy in innovating new technologies. This book thus presented a unique framework of the industrial cluster policy from agglomeration to innovation, which is of practical use. These case studies analysed here focus on identifying policy priority for formation of industrial agglomeration and innovation cluster in emerging economies. The country coverage and thematic commonality are weft and warp of our project. Although knowledge-based cluster issues are important in any developing economies, we have not seen the deeper policy framework in the connection between agglomeration and innovation. To derive a policy framework, empirical evidences on this connection are needed to be accumulated. The policy framework also varies with different development stages and the degree of global and local linkage in each cluster. It is not easy to implement cluster-level policy fostering innovation. We tried to overcome the problem through endogenous innovation in the following ways: (1) comparison among emerging clusters in terms of university–industry linkages (UILs); (2) comparison of organizational change in the face of innovation in each cluster. Through these, we presented both the similarities and differences of the connection between agglomeration and innovation, and reached how to provide evidence-based policy-making in this

Conclusion 359

field. Future studies, based on case studies and econometrics analyses, will focus on identifying the priority of university–industry linkages not only in the context of National Innovation Systems but also Local Innovation Systems at Step II of innovation in the Flowchart Approach to the industrial cluster policy.

References Kuchiki, A. and M. Tsuji (eds) (2005) Industrial Clusters in Asia: Analyses of their Competition and Cooperation. Basingstoke: Palgrave Macmillan. —— (eds) (2008) The Flowchart Approach to Industrial Cluster Policy. Basingstoke: Palgrave Macmillan. Tsuji, M., E. Giovannetti, and M. Kagami (eds) (2007) Industrial Agglomeration and New Technologies. Cheltenham: Edward Elgar.

This page intentionally left blank

Index AFTA (ASEAN Free Trade Agreement), 27 after-sales services, 212–13 Agency for Science, Technology and Research (A*STAR), 63–4, 82, 99 agents, 218, 222–3, 224 agglomeration, 1–6, 20, 118, 174, 357–9 factors encouraging, 6 innovation and, 206 innovation capacity building and, 295–7 labour-based theory of, 291, 293–5 territorial, 167–8, 170–1, 176–7 theory of, 204 workplace training and, 305–8 see also clustering; industrial clusters analysis, 208, 209 Analytical Hierarchical Process (AHP), 232, 233, 238–46 anchor firms, 22–4, 111–12, 270, 282–3 AP Genomics, 76 application software development, 211 Aqua-Terra Supply Co, 101 ASEAN countries future branch expansions in, 29–30 transportation costs between, 27–8 Asian Currency Crisis, 28 Asian-type clusters, 9–10 ASSESSPRO RJ, 166, 167, 184 Association of Singapore Marine Industries (ASMI), 99 Austin, Texas, 25 automobile industry cluster flowchart model patterns and, 17–25 flowchart models, 24–5 in Malaysia, 15–48 prescriptions for Malaysian, 40–4 training and, 292 BAC, see Bioethics Advisory Committee (BAC) Bangalore, India, 9 external linkages of software firms in, 217–23 labour pool in, 218

offshoring-based software clusters in, 204–25 patterns of offshoring in, 209–17 Bangkok, Thailand, industrial clusters in, 290–322 Bio*One Capital, 76 Bioethics Advisory Committee (BAC), 64–5 Bioinformatics Institute, 72 biomedical device start-ups, in Singapore, 75–6 Biomedical Research Council (BMRC), 63–4 biomedical sciences (BMS) cluster clinical research capabilities and, 81–2 development of, 54–63, 65–72 development strategies, 111–14 education and, 83 infrastructure for, 72–3 in Singapore, 54–86 Biomedical Sciences Group, 64 Biopolis, 72 Bioprocessing Technology Institute, 72 biotechnology industry, flowchart models, 24–5 BMRC, see Biomedical Research Council (BMRC) Brazil evolution of oil and gas supply chain in, 333–6 local economic development in, 197–8 oil production in, 328–30 public policy, 198 regional innovation systems (RIS), 333 see also Rio de Janeiro; Rio de Janeiro software cluster Brazilian Development Bank (BNDES), 188 Brazilian Innovation Agency (FINEP), 188 Brenner, Sidney, 74 business incubators, 187–8, 193 business linkages, 52 business management, IT and, 234–5 business process outsourcing (BPO), 210

361

362

Index

Campos Basin, 331–3, 345, 347–9, 350–1 Campus for Research Excellence and Technological Enterprise (CRERATE), 85–6 capacity building, 20, 21, 25, 43 agglomeration and, 295–7 in Brazilian oil and gas sector, 346–9 workplace training and, 290–322 capital human, 45, 293 social, 168 CENPES, 344–5, 346 Centre for Offshore Research and Engineering (CORE), 99 China, 28 Danyang eyewear cluster, 275–9, 283 high-tech zones in, 45 industrial clusters in, 270–84 outsourcing to, 204 specialized markets in, 270–84 Yiwu daily necessities cluster, 273–5, 283 Yuyao mould cluster, 279–82, 283 China Light Industry Association (CLIA), 279 Chiron, 76 CH Offshore, 100 classification of commodities, 274–5 clinical research capabilities, in Singapore, 81–2 clinical trials, 81 clustering, 2–3, 6, 204 external linkages and, 217–23 Flowchart Approach and, 233 innovation and, 206–9, 224–5 offshoring and, 205 regional growth and, 290–1 workplace training and, 295–7 see also industrial clusters codified knowledge, 215 collaboration, private-public, in marine sector, 103–4 collective goods, 168, 169, 172–8, 192–7 Colman, Alan, 74 commodities, classification of, 274–5 competitive environment, 207 consulting firms, 211–12 contract research organizations (CROs), 81

Copeland, Neal, 75 CT Petro, 336–9, 343–4 cultural knowledge, 215–16 customer support, 212–13 daily necessities cluster, 273–5 Danyang eyewear cluster, 275–9, 283 dedicated biotechnology firms (DBFs), in Singapore, 75–80 demand condition, 23–4 deregulation, 38 Diahatsu, 41 domain knowledge, 215–16 Downtown Pole, 183 East Asia industrial transformation in, 1–2 innovation process in, 6–7 see also specific countries econometric method, 23 Economic Development Board (EDB), 99 economic growth, 1, 50–1 education in Singapore, 83 workplace training and, 323n5 educational institutions, 169, 175, 187 Eisai’s Regional Clinical Research Centre, 72 emerging-economy clusters, 207 emerging markets, 272 emerging technologies, 53 endogenous innovation process, 6–7, 117–18, 233–4, 357 entrepreneurs, 207–8 entrepreneurship education programmes, 195–7 environment, 178 European Union, 198 exported-oriented SMEs, 235–7 external agents, 222–3, 224 external economies, 167–9, 171–2 collective goods and, 173–8 external linkages agents, 222–3, 224 of Bangalore software firms, 217–23 innovation and, 205–6, 207–8 inter-firm, 220–1, 224 intrafirm, 218–20 workforce mobility, 221–2, 224

Index 363 eyewear cluster, 275–9 Ezra Holdings, 101 face-to-face interactions, 205, 207, 215–16, 224 financial institutions, 176, 177 FINEP, see Brazilian Innovation Agency (FINEP) FIRJAN, 190, 348 firm-provided training, 290–322 firm size, 142, 146 Flowchart Approach, 3, 15–16, 118, 233–4, 270, 282–3, 292, 357–8 agglomeration, 20 demand condition, 23–4 infrastructure, 21 innovation, 22–3 Malaysian automobile industrial cluster and, 17–25, 40–4, 47–8 flowchart models, 166 differences of, 24–5 patterns of the, 17–25 food industry, 292, 308–21 foreign direct investment (FDI), 1, 38, 54–5, 207 foreign talent, 74–5 Genome Institute of Singapore, 72 global economies, interdependency among, 1 globalization, 173, 174, 204 global technology firms, 210, 216–17 division of labor in, 218–20 intrafirm linkages, 218–20 government Brazilian oil and gas sector and, 336–42 role of, in cluster policy, 10 role of, in developing knowledgebased clustering, 53–4, 112 Gross Expenditure of R&D (GERD), in Singapore, 50–1 Hartwell, Leland, 82 HCL, 210 HDD industry, 292, 301, 308–21 Hewlett Packard (HP), 210 Higashi-Osalka/Ohta, 233, 237–8, 247–57 high-tech sector, external economies and, 173–8

hub-and-spoke industrial district, 16, 291 human capital, 45, 293 human cloning, 65 human resource management policy (HRMP), 323 n1 human resources, 20, 21, 186 in Brazilian oil and gas sector, 342 capacity building and, 290–322 India, 198 offshoring-based software clusters, 209–25 outsourcing to, 204, 213 R&D in, 211 software services exports, 214 Indonesia, 29 industrial agglomeration, 1, 3–4 industrial cluster policy, 25–9, 357–8 industrial clusters, 6 Asian-type, 9–10 automobile, 15–48 in Bangalore, India, 204–25 in Bangkok, Thailand, 290–322 biomedical sciences (BMS) cluster, 54–86 in China, 270–84 competitive environment in, 207 development strategies, 111–14 Flowchart Approach to, 292 innovation and, 118, 132–8 knowledge-based, 51–111 local economic development and, 170–8 machinery, 270 in Malaysia, 15–48 research facilities and, 139, 142 in Rio de Janeiro, 180–1 Rio de Janeiro software cluster, 166–99 Silicon Valley-type, 9–10, 168 in Singapore, 50–114 SMEs and, 117–63 software, 204–25 upgrading, 270–84, 322 see also clustering industrial districts, 3, 16 industrial policy defined, 15–16 Flowchart Approach to, 15–16 flowchart model patterns and, 17–25 in Malaysia, 16–17, 25–9

364 Index industrial upgrading, 124–7, 139–43 firm size and, 142 managerial orientations and, 142–3 research facilities and, 139, 142 results of estimation, 149–57 industrial zones, 22 informal personal networks, 207 information and communication technology (ICT) index of development, 238–46 organizational innovation and, 231–68 promotion of, 232–3 supply chains and, 235–7 information technology industry, flowchart models, 24–5 Infosys Technologies, 210 infrastructure innovation, 186–8 physical, 20, 21, 72–3, 178, 293 innovation, 358 clustering and, 206–9, 224–5 definition and, 234–5 as driver for oil & gas industrial cluster, 326–54 external linkages and, 205–6, 207–8, 217–23 factors encouraging, 6 firm size and, 146 industrial clusters and, 118, 132–8 in Japanese SMEs, 127–30 in knowledge-intensive clusters, 206–9 link between knowledge-based industrial clusters and, 51–4 local, 172–3 local economic development and, 170–8 managerial orientations and, 146–7, 148 new product and services development, 147–8, 162–3 in oil and gas sector, 336–42 organizational, and ICT, 231–68 patents and, 143, 147, 158–61 R&D and, 195 regional innovation systems (RIS), 204, 206–7 research facilities and, 143, 146 Rio de Janeiro software cluster and, 185–97

social construction of, 168 innovation clusters, 169 innovation infrastructure, in Rio de Janeiro, 186–8 innovation mechanism, 4–6 innovation model, 133, 143–8, 158–63 innovation process, 2, 4 in East Asia, 6–7 endogenous, 6, 6–7, 117–18, 357 Flowchart Approach and, 22–3, 118, 233 university-industry linkages and, 44–6 Institute of Bioengineering and Nanotechnology, 72 Institute of Medical Biology, 72, 74 Institute of Molecular and Cell Biology, 72, 74 institutional building, 20, 21 institutional instability, in Malaysia, 29 integration, 208–9 Intel, 219–20 inter-firm cooperation, 2–3 inter-firm linkages, 217, 220–1, 224 International Advisory Council (IAC), 64 International Biomarker Consortium, 82 international maritime centre (IMC), 90–2 international multidisciplinary research, 85–6 international supply chains, 235–7 Internet use, 239, 240 interpretation, 208, 209 inter-sector networks, 53, 82–3 intrafirm linkages, 217, 218–20 investment environment, in ASEAN countries, 32, 34–6 IT cluster, see software and services industry (SSI) IT-enabled services (ITES), 210 IT Hyakusen group, 233, 238, 247–57 IT industry structure and markets, in Rio de Janeiro, 188–92 Ito, Yoshiaki, 74 IT products and services demand for, 190–1 outsourcing of, to India, 204 sources of, 191 IT professionals, 218, 221–2

Index 365 Japan cluster policy in, 10 economy, 117 regional cooperation in, 3 Japan Bank for International Cooperation (JBIC), 29 Japanese auto makers, 41–2 Japanese SMEs, 117–63 characteristics of, 119–24, 130 Higashi-Osalka/Ohta, 237–8 industrial clustering and innovation in, 132–8 innovation model and, 143–8, 158–63 IT Hyakusen group, 238 mail survey of, 118–32 organization innovation in, 231–68 R&D investment, 130 R&D ratio and business performance, 130–2 upgrading and innovation by, 124–30 upgrading estimations, 139–43 upgrading model and, 139–43, 149–57 Japan External Trade Organization (JETRO) survey, 30–6 Jaya Holdings, 100 Jemkins, Nancy, 75 job tenure, 302–3 Keppel FELS, 94, 96, 98, 104–11 knowledge codified, 215 cultural, 215–16 domain, 215–16 tacit, 51, 176, 204 knowledge-based industrial clusters, 51–2 biomedical sciences (BMS) cluster, 54–86 development strategies, 111–14 innovation in, 51–4, 206–9 key processes in development of, 52–3 maritime cluster, 86–111 role of state in developing, 53–4 in Singapore, 54–111 upgrading existing vs. developing new, 54 knowledge creation, 51, 53 knowledge industry, flowchart models, 24–5 knowledge infrastructure, 51–2, 53–4

knowledge spillovers, 204, 293, 294 knowledge transfer, 53, 205, 207 knowledge utilization, 51 knowledge workers, 207–8, 217, 221–2, 224 KS Energy Services, 101 labour markets, 291–5 labour pooling, 204, 205–6, 291 labour shortage, in Malaysia, 28–9, 36, 38 labour-theory based agglomeration, 291, 293–5 labour turnover, 291, 292, 301, 305–8, 321–2 Labroy Marine, 96, 98 Lane, Birgit, 75 life science research, in Singapore, 61–3 Lilly Systems Biology (LSB), 66 Liu, David, 74 living environment, 20, 22 local development, Rio de Janeiro software cluster and, 185–97 local economic development, 170–8, 197–8 local governments in China, 283–4 role of, in cluster policy, 16, 22–3, 25 specialized markets and, 283 local innovation system (LIS), 5, 172–3, 176, 177–8, 357–9 collective goods and, 192–7 Rio de Janeiro, 186–97 localization issues, 32, 33 local markets, 32 local networks, 206 local production systems, 173–4, 177–8, 189, 190–1 logit estimation, 249–51 Lynk Biotechnology, 76 machinery industrial clusters, 270 Malaysia automobile industry cluster in, 15–48 deregulation, 38 domestic market, 37 export possibility, 37 foreign investment and regulation, 38 industrial cluster policy, 25–9 industrial policy, 16–17, 25–9

366 Index Malaysia – continued institutional instability in, 29 investment conditions in, 28–9 ninth Malaysia Plan, 27 situation of, among ASEAN countries, 29–39 supporting industries, 38 Third Industrial Master Plan, 27 transportation costs, 27–8 Malaysian National Car project, 40 managerial orientations, 142–3, 146–7, 148 marine & offshore engineering (M&OE) industry, 90, 94–111 Keppel FELS, 104–11 leading companies, 98 private-public collaborations in, 103–4 R&D in, 99, 102–4 sales revenue, 95–6 SembCorp Marine, 104–11 statistics, 97 Marine and Offshore Technology Centre for Innovation (COI MOT)), 102 Maritime and Port Authority (MPA), 90–2 maritime cluster, in Singapore, 86–111 Maritime Innovation and Technology (MINT) Fund, 92–3 Market Design approach, 283–4 market size, in automobile industry, 42 Marshalian industrial districts, 16, 291 medical research, 66, 72, 82–3 medical technology sector, in Singapore, 55, 57, 61 mergers and acquisitions (M&As), 220–1 Merlion, 66 Microsoft, 219 Mitsubishi Motor Company, 41 MNCs, see multinational corporations (MNCs) mould cluster, 279–82 multinational corporations (MNCs), 1 activities outsources by, 215 division of labor in, 218–20 intrafirm linkages, 217, 218–20 knowledge transfer by, 207 leveraging of, in Singapore, 54–5 onshore-based activities, 215–16 pharmaceutical, in Singapore, 65–72

Nanyang Technological University (NTU), 83, 99 national innovation systems (NISs), 10, 44–6, 357–9 National Organization of the Petroleum Industry (ONIP), 334–5 National Petroleum Agency (ANP), 335–6, 339–42 National University Medical Institutes (NUMI), 84 networks local, 206 personal, 207 ninth Malaysia Plan, 27 NUS, 82, 83, 84, 99 NUS-Duke Graduate Medical School, 84–5 Office of Life Sciences (OLS), 83, 84 offshore and marine engineering cluster innovation capability in, 96–9 in Singapore, 93–111 offshore patents, 103 offshore support services companies, 100–1 offshoring-based software clusters, 204–25 in Bangalore, India, 209–25 challenges of, 216 external linkages and, 205–6 patterns of, 209–17 off-the-job training (OFFJT), 300–3, 312, 317–21, 324n6 oil and gas industrial cluster Campos Basin, 331–3, 347–9, 350–1 capacity building in, 346–9 evolution of Brazilian, 333–6 industry suppliers, 327, 328 innovation in, 336–42 regional innovation systems and, 333 in Rio de Janeiro, 326–54 oil production, in Brazil, 328–30 oil rig platform manufacturers, 94, 98–9 ONIP, see National Organization of the Petroleum Industry (ONIP) on-the-job training, see workplace training Ordered Logit model, 143

Index 367 organizational innovation estimation of, 247–51 factors affecting, 240–1, 243–6, 248–51 at Higashi-Osaka/Ohta, 237–8 ICT and, 231–68 index of ICT development and, 238–46 in IT Hyakusen group, 238 policies encouraging, 255–6 problems with, using ICT, 252–6 outsourcing, 204–5, 213 overseas shops, 236–7 patents, 143, 147, 158–61 PC industry, 292, 301, 308–21 Perodua, 26, 27, 29, 41 personal networks, 207 PETROBRAS, 189, 326–7, 329–30, 332–6, 338–54 pharmaceutical companies, in Singapore, 65–72 pharmaceutical sector, in Singapore, 56, 61 physical infrastructure, 20, 21, 72–3, 178, 293 platform suppliers, 210, 217, 219–20, 224 poaching externality, 293–5, 322, 323 n3 policy measures, Flowchart Approach to, 15–16 Port of Singapore, 90 port sector, 90 private sector actors, 52 probit estimation, 249–51 product development, 147–8, 162–3, 208–9, 211 production process, in high-tech sector, 167–8 production sites, 30–2 production systems, local, 173–4, 177–8, 189–91 professional community, 207 PROMINP, 346–9 Proton, 16–17, 25–6, 41 public policy, 9–10, 169 in Brazil, 198 local economic development and, 171

upgrading industrial clusters and, 322 public R&D institutes (PRIs), 51, 52, 72–3, 74 quality and technical supervision (QTS) department, 277–8 quality control, upgrading, 276–9 REDE PETRO BC, 348–9, 354 n1 Redesoft, 182–3 regional cooperation, 3 regional growth, clustering and, 290–1 regional innovation systems (RIS), 204, 206–7, 333 regional research institutions, 146 research and development (R&D), 10 in Brazilian oil and gas sector, 336–45, 350 business performance and, 130–2 endogenous, 4–6 in India, 211 innovation and, 195 by Japanese SMEs, 130 in marine and offshore engineering industry, 99, 102–4 in Rio de Janeiro, 193 in Singapore, 50–1, 61–3, 66, 84 in Singapore maritime cluster, 92–3 research facilities, 139, 142, 143, 146 research institutions, 169, 177, 193, 221, 270 resource-based industry upgrading, 352–4 see also oil and gas industrial cluster returns to scale, 204 richer hypothesis, 323n4 Rio de Janeiro Campos Basin, 331–3 entrepreneurship education programmes, 195–7 human resources, 186 industrial segments in, 189–90 innovation infrastructure, 186–8 local innovation system, 186–97 oil and gas industrial cluster in, 326–54 public security issues in, 188–9, 194–5 R&D in, 193 regional innovation systems (RIS), 333 SMEs in, 189

368 Index Rio de Janeiro software cluster, 166–99 collective goods and, 192–7 growth of, 178–9 historical evolution of, 179–80 industrial structure, 180–1 industry decline in, 167 innovation and, 185–97 institutional evolution, 182–4 international market and, 194 IT industry structure and markets, 188–92 local development and, 185–97 recent policy actions, 184–5 Rio de Janeiro Technological Park, 183 Rio de Janeiro Technology Network, 333 Rio Knowledge, 183 Rio Software Network, 182–3 S*Bio, 66, 76 satellite industrial districts, 291 Satellite platform, 16 Satyam Computer Services, 210 science parks, 25, 193 scientific institutions, 175 SEBRAE, 347, 348–9, 355 n23 SEBRAE-RJ, 187–8, 190 SembCorp Marine, 94, 96, 98, 104–11 service development, 147–8, 162–3 shipbuilding, 96 ship repairs, 94, 96 Silicon Valley, 6, 168, 205, 217 Silicon Valley-type clusters, 9–10 Singapore, 10 biomedical sciences (BMS) cluster, 54–86 clinical research capabilities, 81–2 cluster development strategies, 111–14 dedicated biotechnology firms in, 75–80 economic growth in, 50–1 industrial cluster development in, 50–114 knowledge-based industrial clusters in, 54–111 maritime cluster, 86–111 medical technology sector, 57, 61 per capita GDP, 50 pharmaceutical companies in, 65–72 pharmaceutical sector, 56, 61 physical infrastructure, 72–3

R&D in, 61–3, 84 R&D infrastructure, 66 translational research, 82–3 universities, 83 university-industry linkages in, 46 Singapore Gastric Cancer Consortium, 85 Singapore Institute for Clinical Sciences, 72, 74 Singapore maritime cluster (SMC), 86–111 competitiveness and, 88–90 development of, 86–93 development strategies, 111–14 growth of, 87 Keppel FELS, 104–11 linkages with economy, 88 offshore and marine engineering cluster, 93–111 R&D and innovation in, 92–3 role of state in developing, 90–3 SembCorp Marine, 104–11 statistics on, 89 Singapore-MIT Alliance (SMA), 85 Singapore-MIT Alliance for Research and Technology (SMART), 85–6 Singapore Technologies Marine, 101 small and medium-sized enterprises (SMEs), 2, 270 in Brazil, 189 export-oriented (case study), 235–7 ICT use and organizational innovation in, 231–68 innovation among, 170–8 innovation model and, 143–8, 158–63 Japanese, 117–63, 231–68 localization of, 169 managerial orientations, 142–3, 146–7, 148 in Rio de Janeiro, 189 territorial concentration of, 173–8 upgrading model and, 139–43, 149–57 social capital, 168 social construction of innovation, 168 social networks, 177 software and services industry (SSI) collective goods and, 192–7 external economies and, 167–8 external linkages and, 217–23 face-to-face interactions in, 207 nature of, 210–13 Rio de Janeiro, 166–99

Index 369 software clusters, offshoring-based, 204–25 software development, 215–17, 219–20 software industry, division of labor in, 215–17 software services exports, 214 software solutions and consultancy, 211–12 software use, 239, 240 South Rio Pole, 183–4 spatial economic theory, 16 specialized markets case studies, 273–82 in China, 270–84 features of, 271–2 Special Participation clause, 339–42, 345 SPRING Singapore, 102 start-up companies, 195 state role of, in developing knowledgebased clustering, 53–4 role of, in Singapore maritime cluster, 90–3 stem cell research, 65 suppliers, 175–6 supply chain, 235–7, 333–6 Swain, Judith, 75 Swiber, 100 Swissco International, 100 tacit knowledge, 51, 176, 204 Taiwan, 222 Tata Consultancy Services (TCS), 210 TCS, 218, 221, 222 technological districts, 168–9 Technological Innovation Programme (TIP), 102 Temasek Life Sciences Laboratory (TLL), 84 territorial agglomeration, 167–8, 170–1, 176–7 Texas Instruments (TI), 210, 219 Thailand, 16, 29 industrial clusters in, 290–322 Third Industrial Master Plan, 27 training, see workplace training translational research, 66, 72, 82–5

transportation costs, between countries in ASEAN, 27–8 trust, 177 Ullrich, Axel, 75 universities, 51, 52, 177, 186, 187, 270 in Brazil, 193, 333 partnerships with, 221 in Singapore, 83 university-industry linkages (UILs), 44–6 upgrading policy recommendations for, 322 quality control, 276–9 resource-based industry, 352–4 role of specialized markets in, 270–84 value chains, 280–2 upgrading model, 133, 139–43 results of estimation, 149–57 value chains, 280–2 venture capital, 75–6 Vietnam, 29 Wenk, Markus, 75 West Clinic’s Excellence Cancer Centre, 72 Wipro Technologies, 210 workforce mobility, 221–2, 224 workplace training capacity building and, 290–322 clustering firms and, 295–7 data on, 297–304 industry comparisons, 305–21 poaching externality and, 293–5, 322 turnover rates and, 305–8, 308–21, 321–2 Ying, Jackie, 75 Yiwu China Commodities City, 273–5, 283 Yuyao Market, 279–82 Yuyao mould cluster, 279–82, 283 Zhejiang China Commodities City Group (ZCCC Group), 273 Zhejiang Moulds Production Center (ZMPC), 279